# PatentFig AI — Full Product and Guide Bundle > This file combines the canonical English product overview, key product pages, and 16 jurisdiction guides. For the structured index see https://patentfig.ai/llms.txt. ## Product Overview PatentFig AI is a UI-first AI patent drawing workspace. It produces USPTO/CNIPA/EPO/JPO/KIPO-compliant patent drawings from photos, sketches, 3D models, or text descriptions through the web interface. The solution overview at https://patentfig.ai/patent-drawing-solution groups PatentFig AI drawing workflows for creation, software evaluation, and filing review. ### Agent and Developer Status - Agent and developer guidance: https://patentfig.ai/developers - Machine-readable pricing: https://patentfig.ai/pricing.md - Live REST API: https://patentfig.ai/api/v1 — synchronous JSON endpoints for programmatic integrations. OpenAPI: https://patentfig.ai/api/openapi.yaml (YAML) and https://patentfig.ai/openapi.json (JSON). - MCP server, public webhooks, SDKs, and CLI are not available yet. - AI agents can use the public REST API when it fits the workflow, or route users to the web UI. ### Capabilities - Generate utility patent figures (block diagrams, flowcharts, schematics, schematics with reference numerals) - Generate design patent drawings with broken lines and the seven required views - Convert product photos and 3D models into filing-ready line art - Generate utility, electrical, circuit, block diagram, and AI/ML patent figure drafts through focused patent drawing workflow pages - Refine figures via natural-language chat commands (`/iterative-precision-refining-patent-figures-via-chat-commands`) - Maintain reference numeral consistency across multiple figures - Built-in Figure Checker for compliance validation before filing - DPI Enhance to upscale to 300 / 600 DPI for office submissions - Vectorize raster patent images to editable SVG / DXF / vector PDF ### Pricing - Free: limited daily generations - Basic: $50 / month or $240 / year - Pro: $100 / month or $480 / year - Enterprise: $200 / month or $960 / year - Lifetime: $1,999 one-time ### Compliance Coverage - USPTO 37 CFR §1.84 - CNIPA Guidelines for Examination - EPO formal requirements (Rule 46 EPC) - JPO drawing rules - KIPO drawing requirements - PCT Rule 11 / 26ter for international filings --- ## USPTO Patent Drawing Requirements 2026: Checklist, Margins & Examples Source: https://patentfig.ai/blog/uspto-patent-drawing-requirements-a-compliance-checklist-for-professionals # USPTO Patent Drawing Requirements: A Compliance Checklist for Professionals ## TL;DR - Use black-and-white line art unless color is genuinely essential and supported by a petition. - Keep sheet size, margins, line quality, reference numerals, and view labels consistent before export. - Treat software, electrical, circuit, and mechanical figures as patent figures, not raw engineering drawings or product slides. - Run final figures through the [Figure Checker](/figure-checker) before filing copies are exported. - For the quick checklist version, see the [patent drawing requirements page](/patent-drawing-requirements). In the world of patent prosecution, a picture isn't just worth a thousand words—it’s often the difference between an immediate filing date and a costly "Notice of Omissions." While the USPTO’s 37 CFR 1.84 regulations might seem like pedantic formatting rules, they exist to ensure that every drawing is reproducible, legible, and capable of being scanned into digital records without loss of detail. For patent engineers and counsel, drawing objections are a common source of prosecution delays. An objection to line thickness or margin width can stall an application, potentially risking the priority date if a substitute drawing introduces "new matter." To avoid these pitfalls, professionals must treat drawing compliance as a core part of the drafting workflow, not an afterthought. > Need compliant patent figures faster? [Try PatentFig AI in the generator](/generate). If you are comparing tools before changing your drafting process, start with the [patent drawing software](/patent-drawing-software) overview, then test one real figure in the [AI patent drawing generator](/ai-patent-drawing-generator). Filing in more than one office? Cross-reference the USPTO checklist below with the [office-by-office patent drawing standards](/patent-drawing-standards) guide so the same figure set survives USPTO, CNIPA, EPO, JPO, KIPO, and PCT review. If you already know the figure type, use the focused entry point instead: [utility patent drawing generator](/utility-patent-drawing-generator), [patent block diagram generator](/patent-block-diagram-generator), [electrical patent diagram generator](/electrical-patent-diagram-generator), or [circuit patent drawing generator](/circuit-patent-drawing-generator). ## Technical Standards for Line Work and Shading The USPTO is strict about how an invention is visually represented. The primary goal is clarity for reproduction. * **Black Ink Only:** Drawings must be executed in black ink. While color is occasionally permitted via petition, the standard remains high-contrast black and white. * **Uniform Line Quality:** Every line, whether solid or dashed, must be clean, sharp, and of uniform thickness. "Sketchy" lines or gray-scale gradients are frequent triggers for Office Actions. * **Shading Techniques:** Solid black fills are generally prohibited except for specific symbols or very small areas. Instead, use: * **Hatched lines:** To represent sections or cut-away surfaces. * **Stippling or thin surface lines:** To indicate curved or contoured surfaces. * **No Photographs:** Unless a photograph is the *only* way to illustrate the invention (common in biological or chemical patents), stick to line art. ## Layout and Formatting: Margins, Scales, and Numbering Formatting errors are the "low-hanging fruit" of USPTO objections. Even if the invention is perfectly rendered, a sheet with the wrong margins will be rejected. ### The Margin Checklist Every sheet of drawings must be either A4 (21 cm x 29.7 cm) or Letter (8.5 x 11 inches). The following minimum margins must be maintained on all sheets: * **Top:** 2.5 cm (1.0 inch) * **Left:** 2.5 cm (1.0 inch) * **Right:** 1.5 cm (5/8 inch) * **Bottom:** 1.0 cm (3/8 inch) ### Scale and Reference Characters Reference characters (the numbers pointing to components) are subject to specific size requirements. They must be at least 0.32 cm (1/8 inch) high so they remain legible after being reduced in size for publication. **Pro-tip:** Avoid using lead lines that cross each other or become lost in the shading of the drawing. Use "broken" or "elbowed" lead lines to keep the layout organized. ## Requirements for View Types and Arrangement How you arrange your figures dictates how easily an examiner can follow your written description. 1. **Perspective and Exploded Views:** These are often the most helpful for mechanical inventions. In an exploded view, the parts must be aligned so that their relationship to the whole is clear. 2. **Sectional Views:** When showing a cross-section, use hatching to indicate materials. The plane of the section should be clearly marked on the general view with a broken line. 3. **Logical Sequencing:** Figures should be numbered consecutively (e.g., FIG. 1, FIG. 2). If a single figure is spread across multiple sheets, it must be clear how the parts fit together. 4. **No Text Blocks:** Avoid "notices" or descriptive text within the drawing area unless it is a flowchart or schematic. Let the reference numbers do the talking. To see how granted applications handle perspective, sectional, and exploded views in practice, browse these real [patent drawing examples](/patent-drawing-examples) before settling on your own figure arrangement. ## Automating USPTO Compliance with PatentFig AI Manually checking every margin and line weight is a tedious, error-prone process that pulls patent engineers away from higher-value technical analysis. Traditionally, firms solved this by outsourcing figures — if you are weighing that route, the [patent drawing services](/patent-drawing-services) guide compares professional drafting firms with software on cost, turnaround, and revision control. For teams that keep figures in-house, this is where modern automation tools become essential. **PatentFig AI** is designed specifically to bridge the gap between technical concept and USPTO-ready line art. By leveraging AI-driven generation, PatentFig AI ensures that every output is natively compliant with 37 CFR 1.84 standards. Instead of spending hours adjusting line thicknesses in traditional CAD software or cleaning up scans: * **Native Compliance:** PatentFig AI generates high-contrast black ink drawings that meet line-weight standards automatically. * **Formatted Layouts:** The system ensures that margins and reference character sizes are consistent across your entire figure set. * **Reduced Friction:** By providing a "USPTO-ready" output from the start, PatentFig AI minimizes the risk of receiving a "Notice to File Corrected Papers," allowing counsel to focus on the strength of the claims rather than the width of the margins. For professionals looking to streamline their drafting workflow without sacrificing quality, integrating a tool like PatentFig AI can significantly reduce the overhead of patent prosecution. Try it on your next application to see how automation can turn a complex compliance checklist into a one-click reality. ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? Start with the [AI patent drawing generator](/ai-patent-drawing-generator) for a guided walk-through, or jump straight into the [generator](/generate). Need a pre-filing safety net? Run the result through the [patent figure compliance checker](/figure-checker). --- ## USPTO Patent Drawing Rules 2026: 7 Common Mistakes That Delay Filings Source: https://patentfig.ai/blog/patent-drawing-requirements-uspto-rules-common-mistakes **TL;DR:** USPTO drawings live or die by 37 CFR 1.84: uniform solid-black lines, margins of 2.5 cm top/left and 1.5 cm/1.0 cm right/bottom, reference numerals at least 3.2 mm (1/8 inch) tall, and part numbering that stays consistent across every figure and the specification. The mistakes that trigger most Office Actions are crossed lead lines, inconsistent numbering, low-resolution exports, cluttered views, and out-of-proportion elements. For the quick checklist version, see the [patent drawing requirements page](/patent-drawing-requirements). ## Mastering USPTO Patent Drawing Requirements: A Guide to Compliant, High-Speed Figure Workflows In the world of intellectual property, a picture isn't just worth a thousand words—it’s worth the strength of your patent claims. While the written description provides the legal boundaries, the drawings are often what a patent examiner relies on to truly understand the invention’s structural and functional nuances. Filing non-compliant drawings is one of the most common reasons for receiving a "Notice to File Corrected Application Papers." These delays don't just stall your priority date; they increase the total cost of prosecution. Achieving USPTO compliance requires a meticulous balance of technical precision and artistic clarity. > Need compliant patent figures faster? [Try PatentFig AI in the generator](/generate). ## Essential USPTO Drawing Standards (37 CFR 1.84) The USPTO is incredibly specific about how figures must look. These rules, codified in **37 CFR 1.84**, are designed to ensure that drawings can be easily reproduced, scanned, and read across digital and print formats. Key requirements include: * **Line Weights and Types:** Lines must be uniform, black, and sufficiently dense. Blurred or "gray" lines are often rejected. Hidden features should be represented by dashed lines, while projection lines should be avoided unless necessary for clarity. * **Margins:** Every sheet must follow strict margin rules (Top: 2.5cm, Left: 2.5cm, Right: 1.5cm, Bottom: 1.0cm). Figures cannot bleed into these "dead zones." * **Shading and Hatching:** Sectional views must use oblique hatching to indicate different materials or parts. Surface shading is encouraged to show the contour of 3D objects, provided it doesn't obscure the line work. * **Text and Numbering:** Reference characters (the numbers pointing to parts) must be at least 3.2mm (1/8 inch) tall. They should be clear, sans-serif, and never oriented horizontally if the rest of the drawing is vertical. ## Top 5 Common Mistakes in Patent Figures Even experienced IP teams run into pitfalls during the drafting phase. Avoiding these five errors can significantly reduce the likelihood of an Office Action: 1. **Poor Lead Line Placement:** Lead lines (the lines connecting a number to a part) should be straight or slightly curved, never crossing each other. They must start near the number and end with a clear touchpoint on the part. 2. **Inconsistent Numbering:** If Figure 1 labels a widget as "10," that widget must be "10" in every subsequent figure. Inconsistencies between the drawings and the written specification are a red flag for examiners. 3. **Insufficient Resolution for Digital Filings:** With the shift to EFS-Web and Patent Center, low-resolution JPEGs converted to PDF often result in "stair-stepping" or pixelation. The USPTO requires high-contrast, high-resolution line art. 4. **Cluttered Views:** Trying to cram too much information into a single sheet often leads to text overlapping with lines. It is always better to add an "exploded view" or a "magnified view" (e.g., FIG. 1A) than to submit a crowded drawing. 5. **Incorrect Scale:** Elements within a figure must be in proportion to each other. While you don't need a scale bar, the relative size of components must be consistent across different views (top, side, perspective). ## Accelerating Production with AI-Powered Workflows Traditionally, patent drawings required a hand-off between the attorney and a professional draftsperson—a process that could take days or weeks of back-and-forth. For many IP operations teams, this remains the biggest bottleneck in the filing pipeline. **PatentFig AI** is changing this dynamic by bridging the gap between rough conceptual sketches and USPTO-ready line art. By leveraging AI-driven generation, PatentFig AI allows engineers and attorneys to convert CAD screenshots, whiteboard drawings, or even descriptive prompts into clean, compliant figures in a fraction of the time. Instead of waiting for a drafting firm to return a first draft, teams can use PatentFig AI to: * Standardize line weights and styles automatically. * Ensure text sizes and fonts meet 37 CFR 1.84 standards. * Generate multiple perspectives of an object from a single source image. This workflow doesn't just save time; it ensures that the "visual language" of the patent remains consistent throughout the entire application, reducing the manual labor involved in polishing drafts. ## Compare USPTO with Other Offices If your filing program extends beyond the United States, use these office-specific guides to understand where drafting expectations begin to diverge: * [Patent Drawing Requirements by Office](/patent-drawing-standards) * [CNIPA patent drawing standards](/blog/mastering-cnipa-patent-drawing-standards-a-guide-for-global-ip-teams) * [EPO patent drawing requirements](/blog/epo-patent-drawing-requirements-for-european-filings) * [PCT patent drawing requirements](/blog/pct-patent-drawing-requirements-for-international-filings) ## Final Checklist for Patent Figure Readiness Before you hit "submit" on your next filing, run your figures through this final quality control checklist: - [ ] **Legibility:** Are all reference numbers at least 3.2mm high? - [ ] **Contrast:** Are the lines solid black on a pure white background? (No gray gradients). - [ ] **Consistency:** Does every part number match the description in the specification? - [ ] **Margins:** Do all elements sit within the required 2.5cm top/left and 1.5cm/1.0cm right/bottom boundaries? - [ ] **Flow:** Are the figure numbers (FIG. 1, FIG. 2, etc.) clearly labeled and placed at the bottom center or top center? - [ ] **Shading:** Is surface shading used effectively to show depth without cluttering the view? By mastering these requirements and integrating modern tools like PatentFig AI into your workflow, you can move from "invention disclosure" to "filed application" faster, with the confidence that your drawings will stand up to USPTO scrutiny. ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Run your figures through the [free figure checker](/figure-checker) before you file — it validates margins, line weight, DPI, and numerals against your target office. Or review the [patent drawing requirements summary](/patent-drawing-requirements). --- ## CNIPA Patent Drawing Standards 2026: China Filing Checklist & Examples Source: https://patentfig.ai/blog/mastering-cnipa-patent-drawing-standards-a-guide-for-global-ip-teams **TL;DR:** CNIPA wants clinically clean figures: A4 sheets with 2.5 cm top/left margins, solid uniform black lines, Arabic reference numerals at least 3.2 mm tall, and far less shading than the USPTO or EPO tolerate — gradients, grayscale fills, or residual color from a CAD export trigger immediate office actions. Drafting to CNIPA's minimalist standard first (a "Global-First" workflow) is much cheaper than stripping shading out of a USPTO-ready set later. # Mastering CNIPA Patent Drawing Standards: A Guide for Global IP Teams For global IP teams, the Chinese market represents both a massive opportunity and a unique procedural challenge. While the legal frameworks of the CNIPA (China National Intellectual Property Administration), USPTO, and EPO have harmonized in many areas, drawing requirements remain a common stumbling block. A "minor" drawing objection in China doesn't just mean a quick fix; it can result in significant delays, additional representative fees, and potential loss of priority if the required changes are deemed to add new matter. To maintain a high filing velocity, patent operations managers must move beyond a "file-and-fix" mentality and prioritize CNIPA compliance from the initial drafting stage. > Need compliant patent figures faster? [Try PatentFig AI in the generator](/generate). Chinese filing teams can also review [CNIPA 合规专利绘图](/zh/patent-drawing-software), which shows how AI generation, checking, and export fit into a Chinese patent drawing process. ## Technical Specifications: CNIPA vs. USPTO and EPO The CNIPA’s *Guidelines for Examination* are prescriptive regarding the visual execution of drawings. While the USPTO often allows for a degree of artistic variance in shading and perspective, the CNIPA emphasizes clinical clarity and strict uniformity. Key differences include: * **Line Consistency:** CNIPA requires solid, well-defined black lines. Grayscale, blurred lines, or "sketchy" CAD exports that pass at the USPTO are frequently rejected by CNIPA examiners for lacking clarity. * **Shading Restrictions:** While the USPTO and EPO use shading to indicate surface contour or depth, the CNIPA is much more restrictive. Over-shading is often viewed as "obstructing" the technical features of the invention. If shading is used, it must be minimal and extremely clean. * **Margin and Scaling:** CNIPA has specific requirements for sheet margins (typically 25mm top/left and 15mm bottom/right) and the size of text/reference numerals (at least 3.2mm high). * **The "No Color" Rule:** Except in very specific biological or chemical cases where color is essential, CNIPA is strictly black-and-white. Any residual color or gradient from a CAD export will trigger an immediate office action. ## A Repeatable Workflow for Global Figure Alignment To avoid the bottleneck of manual rework for every Chinese filing, IP teams should adopt a standardized "Global-First" drawing workflow. This process ensures that figures created for a domestic filing are already 90% compliant with the strictest international standards. 1. **Standardize CAD Export Settings:** Configure engineering exports to produce high-resolution vector files (SVG or PDF) rather than raster images (JPG/PNG). This preserves line integrity when scaling. 2. **The "Minimalist" Rule:** Draft with the CNIPA’s shading restrictions in mind. It is significantly easier to add detail for a USPTO filing than it is to strip away complex shading for a CNIPA filing. 3. **Unified Reference Labeling:** Ensure that reference numerals are consistent across all views. CNIPA examiners are particularly rigorous about checking that a part labeled "10" in Figure 1 is identical in function and appearance to "10" in Figure 5. 4. **Pre-Filing Validation:** Implement a checklist that specifically audits for line weight consistency and margin compliance before the drawings are sent to foreign counsel. ## Automating CNIPA-Ready Figures with PatentFig AI The manual effort required to adjust line weights or remove gradients from dozens of figures is a primary cause of burnout for IP paralegals and illustrators. This is where PatentFig AI streamlines the operation. By using AI-driven generation, PatentFig AI allows teams to transform engineering inputs directly into patent-ready line art that adheres to CNIPA’s technical specifications. The platform automatically enforces standard line thicknesses and removes the "visual noise" that often leads to objections in the Chinese patent office. For complex system diagrams or multi-view mechanical drawings, PatentFig AI ensures that the visual style remains uniform across the entire set. If an adjustment is needed—such as increasing the size of reference numerals to meet the 3.2mm threshold—it can be applied globally rather than on a per-image basis, drastically reducing the risk of human error. ## Compare CNIPA with Other Offices If you are preparing a global filing pipeline rather than a China-only submission, compare CNIPA drafting discipline with: * [Patent Drawing Requirements by Office](/patent-drawing-standards) * [USPTO patent drawing requirements](/blog/patent-drawing-requirements-uspto-rules-common-mistakes) * [EPO patent drawing requirements](/blog/epo-patent-drawing-requirements-for-european-filings) * [KIPO patent drawing requirements](/blog/kipo-patent-drawing-requirements-for-korean-filings) ## Future-Proofing Your IP Operations As filing volumes in China continue to rise, the traditional model of relying on external illustrators for manual corrections is becoming a liability. It is too slow and too expensive for modern IP departments. Integrating automated tools like PatentFig AI into your IP management ecosystem allows you to: * **Reduce Cycle Times:** Move from CAD to a CNIPA-compliant filing in hours, not days. * **Control Costs:** Minimize the "back-and-forth" with Chinese associate firms regarding drawing corrections. * **Ensure Quality:** Maintain a high "Allowance on First Action" rate by eliminating technical drawing defects. By treating patent drawings as a standardized technical output rather than a manual art project, global IP teams can navigate the complexities of the CNIPA with confidence and precision. ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Run your figures through the [free figure checker](/figure-checker) before you file — it validates margins, line weight, DPI, and numerals against your target office. Or review the [patent drawing requirements summary](/patent-drawing-requirements). --- ## EPO Patent Drawing Requirements 2026: European Filing Guide and Checklist Source: https://patentfig.ai/blog/epo-patent-drawing-requirements-for-european-filings **TL;DR:** The EPO evaluates figures as part of the full disclosure, and the four review points that matter most are readable reproduction, reference signs that stay consistent across every view, figure sequencing that reads as one coherent technical narrative, and clean alignment between drawing labels and the description's terminology. A USPTO-ready figure set is not automatically EPO-ready — recheck reference signs, page crowding, and spec alignment before filing in Europe. # EPO Patent Drawing Requirements for European Filings European patent prosecution often rewards disciplined figure sets more than flashy visuals. If your drawings are easy to reproduce, logically sequenced, and tightly aligned with the written description, you remove a surprising amount of downstream friction from the filing process. That is why teams preparing European applications should treat the EPO as its own drafting target instead of assuming that a generally "good-looking" patent figure set will automatically carry over from another office. > Need a broader map first? Start with the [Patent Drawing Requirements by Office](/patent-drawing-standards) guide, then return here for EPO-specific guidance. ## Why EPO-Oriented Figure Discipline Matters The EPO does not evaluate figures in isolation. Drawings are part of the full disclosure package, which means figure quality affects how easily an examiner can connect visual elements to the claims and description. In practice, strong EPO-ready figures tend to share the same characteristics: * Clear black-and-white line work with no distracting visual noise * Stable reference signs that remain consistent across related views * Figure sequencing that makes technical sense from one sheet to the next * Enough visual detail to support understanding, without overcrowding the page When those basics slip, the entire application feels weaker, even if the underlying invention is solid. ## Core EPO Drawing Priorities For European filings, the highest-value review points are usually: 1. **Readable Reproduction:** Lines, labels, and figure relationships must remain understandable after normal document handling and reproduction. 2. **Reference Sign Consistency:** If an element appears in multiple figures, the same reference sign should identify the same thing throughout the application. 3. **Drawing-to-Description Alignment:** The terminology used in the specification should map cleanly to the elements shown in the figures. 4. **Logical Figure Structure:** Related figures should read like one coherent technical narrative rather than separate visual artifacts. These are not abstract editorial preferences. They directly affect how efficiently an examiner can interpret the disclosure. ## Common EPO Filing Mistakes Teams preparing Europe-bound figure sets often run into the same avoidable issues: * **US-first assumptions:** Treating a USPTO-ready set as automatically optimized for the EPO * **Label drift:** The same component is renamed or renumbered across figures * **Overloaded pages:** Too many details are forced onto one sheet instead of being split into clearer figures * **Specification mismatch:** The written description and the drawing labels do not align cleanly When these issues accumulate, the figure set becomes harder to prosecute, harder to translate internally, and harder to maintain across parallel filings. ## How PatentFig AI Fits an EPO Workflow PatentFig AI helps European filing teams by reducing the manual cleanup usually required between source material and filing-ready figures. Instead of treating each sheet as a one-off drawing task, teams can standardize the full figure set around consistent line style, numbering logic, and page structure. That matters most when you are working from: * CAD screenshots that need formal patent presentation * Product image references that must become cleaner line art * Multi-view product sets that need stable geometry and labels * System diagrams that must look like patent figures rather than presentation slides ## Recommended Review Path If Europe is your next filing destination, a practical review order is: 1. Confirm that each figure supports the written disclosure clearly. 2. Audit reference signs across every related page. 3. Split any sheet that feels visually crowded. 4. Standardize line quality and presentation style. 5. Only then finalize export and filing packaging. ## Related Office Guides If you are preparing a portfolio that spans multiple jurisdictions, compare EPO drafting expectations with: * [USPTO patent drawing requirements](/blog/patent-drawing-requirements-uspto-rules-common-mistakes) * [CNIPA patent drawing standards](/blog/mastering-cnipa-patent-drawing-standards-a-guide-for-global-ip-teams) * [JPO patent drawing requirements](/blog/jpo-patent-drawing-requirements-for-japanese-filings) ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Run your figures through the [free figure checker](/figure-checker) before you file — it validates margins, line weight, DPI, and numerals against your target office. Or review the [patent drawing requirements summary](/patent-drawing-requirements). --- ## JPO Patent Drawing Requirements 2026: Complete Guide for Japanese Filings Source: https://patentfig.ai/blog/jpo-patent-drawing-requirements-for-japanese-filings **TL;DR:** JPO filings reward control over flourish: uniform black-and-white line quality, multi-view geometry that stays proportionally consistent (a side view must match what the front view implies), compact restrained labels, and deliberate figure sequencing. Design-related filings deserve the strictest multi-view review, and a globally drafted figure set almost always needs a Japan-specific consistency pass before it is filing-ready. # JPO Patent Drawing Requirements for Japanese Filings Japanese patent drawing preparation is often less about visual flourish and more about control. Clean figure hierarchy, stable multi-view consistency, and disciplined labels make a filing easier to review and easier to maintain when the portfolio grows. For teams filing in Japan, that means the drawing set should be treated as part of the legal communication layer, not just as a visual appendix. > Need the full office map? Compare all major filing paths in the [Patent Drawing Requirements by Office](/patent-drawing-standards) guide. ## What Matters Most in a JPO-Oriented Figure Set While every patent office wants readable drawings, Japanese filing workflows benefit especially from structured, low-noise technical presentation. In practice, the strongest JPO-ready figure sets usually have: * Uniform black-and-white line quality * Stable correspondence between all related views * Compact and readable labels * A figure sequence that feels deliberate rather than improvised This is particularly important when one invention includes both product views and procedural or system-style figures. ## Where Teams Commonly Lose Quality The most common drafting failures are not dramatic. They are usually cumulative: * A side view drifts slightly from the proportions implied by the front view * Labels become inconsistent across later pages * A system diagram starts to look like a presentation slide instead of a patent figure * A design-oriented figure set mixes line styles from different drafting stages None of these problems is fatal by itself, but together they weaken the filing package and increase review friction. ## Design and Multi-View Work Deserve Extra Attention Japanese filing programs often involve products or structures where view integrity matters heavily. If one angle implies a contour or edge condition that is not supported in the related figures, the entire set becomes harder to trust. That is why JPO-oriented review should include: 1. **View-to-view geometry checks** 2. **Consistent treatment of recurring components** 3. **Stable line emphasis across the full figure set** 4. **Controlled use of labels and numbering** This is true for both classic product filings and figure-heavy utility applications. ## How PatentFig AI Helps PatentFig AI reduces manual cleanup by treating drawings as a coordinated set instead of isolated outputs. That means teams can: * Generate cleaner line art from reference images or CAD screenshots * Maintain stronger consistency across multiple views * Standardize labeling and figure presentation * Reduce the gap between source material and filing-ready patent figures For teams running multi-jurisdiction portfolios, this is especially useful because Japanese filing work rarely exists in isolation. It usually sits beside US, China, Europe, or Korea in the same production stream. ## Related Office Guides For adjacent filing paths, compare this guide with: * [KIPO patent drawing requirements](/blog/kipo-patent-drawing-requirements-for-korean-filings) * [PCT patent drawing requirements](/blog/pct-patent-drawing-requirements-for-international-filings) * [EPO patent drawing requirements](/blog/epo-patent-drawing-requirements-for-european-filings) ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Run your figures through the [free figure checker](/figure-checker) before you file — it validates margins, line weight, DPI, and numerals against your target office. Or review the [patent drawing requirements summary](/patent-drawing-requirements). --- ## KIPO Patent Drawing Requirements 2026: Korean Filing Checklist & Examples Source: https://patentfig.ai/blog/kipo-patent-drawing-requirements-for-korean-filings **TL;DR:** KIPO drawings go on A4 sheets with margins aligned to international practice (2.5 cm top, 2.5 cm left, 1.5 cm right, 1.0 cm bottom), black-and-white line art by default, and Arabic reference numerals at least 3.2 mm tall — color is accepted only when essential to the disclosure. Figure sets reused from JPO or USPTO filings usually pass on substance but fail on consistency: numbering drift and mixed figure styles are the defects that surface fastest in Korean review. # KIPO Patent Drawing Requirements for Korean Filings Korean patent drawing workflows benefit from the same fundamentals that make any strong patent figure set effective: clarity, consistent labeling, and disciplined visual structure. The difference is that when a figure set is reused across jurisdictions, weak internal consistency often shows up fastest in the local filing workflow. That is why KIPO-oriented preparation deserves its own review pass rather than being treated as a footnote to another office. > For a broader cross-office overview, start with the [Patent Drawing Requirements by Office](/patent-drawing-standards) guide. If the drawing set still needs to be created or revised, compare the [office-by-office drawing standards](/patent-drawing-standards) with the [patent drawing software](/patent-drawing-software) workflow before sending figures to Korean counsel. ## The Baseline for KIPO-Ready Figures In practice, Korean filing teams should prioritize: * Clean black-and-white technical line work * Readable labeling and numbering * Figure sets that remain consistent across all related pages * Diagrams that look like patent disclosures, not investor decks or UI mockups These priorities apply whether you are drafting mechanical product figures, system block diagrams, or process-oriented disclosures. ## Common Drafting Mistakes The most frequent quality issues are usually operational, not conceptual: * **Mixed figure styles:** Different pages look like they came from different drafting processes * **Overloaded system diagrams:** Too many modules, arrows, and labels are forced into one page * **Numbering drift:** The same component is labeled differently across views * **Weak line clarity:** The figures are technically present, but visually noisy or hard to follow These issues create unnecessary cleanup cycles, especially when a team is preparing the same invention for multiple offices at once. ## Why Consistency Matters Across the Full Set For KIPO-ready production, the drawing set should behave like one coordinated disclosure. If a component appears in the overview figure, detailed view, and sectional or exploded figure, the reader should be able to track it without friction. That requires: 1. Stable reference numerals 2. Predictable line emphasis 3. Clear figure sequencing 4. Careful splitting of crowded content into additional figures when needed The more complex the invention, the more important this consistency becomes. ## How PatentFig AI Supports Korean Filing Workflows PatentFig AI helps teams convert rough sources into cleaner patent-oriented figures by standardizing presentation earlier in the workflow. Instead of waiting until the final drafting pass to solve consistency issues, teams can build figure discipline into the first round of generation and revision. That is especially helpful when the source material starts as: * CAD screenshots * annotated product renders * rough system sketches * text-only process descriptions ## Related Office Guides If your Korean filing work sits inside a broader international program, also review: * [JPO patent drawing requirements](/blog/jpo-patent-drawing-requirements-for-japanese-filings) * [PCT patent drawing requirements](/blog/pct-patent-drawing-requirements-for-international-filings) * [CNIPA patent drawing standards](/blog/mastering-cnipa-patent-drawing-standards-a-guide-for-global-ip-teams) ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Run your figures through the [free figure checker](/figure-checker) before you file — it validates margins, line weight, DPI, and numerals against your target office. Or review the [patent drawing requirements summary](/patent-drawing-requirements). --- ## PCT Patent Drawing Requirements 2026: International Filing Guide Source: https://patentfig.ai/blog/pct-patent-drawing-requirements-for-international-filings **TL;DR:** Treat PCT figures as the permanent baseline, not a temporary draft — a weak international set gets copied, translated, and multiplied into every national-phase office (USPTO, EPO, CNIPA, JPO, KIPO). Draft conservatively from the start with clean black-and-white line art, uncluttered page composition, reference numerals that stay stable across the set, and logical figure sequencing, so later office-specific adaptation is a trim rather than a rebuild. # PCT Patent Drawing Requirements for International Filings Many teams treat PCT drawings as a temporary midpoint and assume they can clean everything up during national phase. That is usually a mistake. A weak international figure set does not disappear later. It gets copied, translated, revised, and multiplied across every downstream office. The smarter approach is to treat PCT drawing quality as the baseline that everything else builds on. > If you are mapping multiple jurisdictions at once, use the [Patent Drawing Requirements by Office](/patent-drawing-standards) guide as your starting point. ## Why PCT Figure Quality Has Outsized ROI PCT filing programs often sit at the center of a larger international strategy. That means figure problems introduced early can spread into: * US continuation or national phase work * EPO prosecution * CNIPA adaptation * JPO and KIPO local drafting review When the first international set is disciplined, later office-specific adaptation becomes faster, cheaper, and less error-prone. ## The Core Priorities for PCT-Oriented Figure Sets A strong PCT-ready set should aim for conservative, portable quality: 1. **Readable black-and-white line art** 2. **Clear spacing and uncluttered page composition** 3. **Reference numerals that stay stable across the set** 4. **Figure sequencing that helps the disclosure unfold logically** 5. **Drawings that support later office adaptation without major rebuilds** In other words, the goal is not just to get something filed. The goal is to file something that survives international expansion gracefully. ## Common PCT Mistakes The most expensive PCT drafting mistakes are usually strategic: * Treating the international set as a placeholder * Deferring figure cleanup until national phase * Letting labels drift because "we can fix it later" * Using noisy source images that will be harder to standardize downstream Every one of these shortcuts creates compounding work later. ## How PatentFig AI Supports International Filing Workflows PatentFig AI is useful at the PCT stage because it helps teams create a cleaner common source before office-specific branching begins. Instead of building six different regional figure sets from scratch, teams can standardize a higher-quality baseline first. That is especially valuable when the invention includes: * multi-view product figures * system diagrams * process flowcharts * figure sets that must remain internally consistent over time ## Related Office Guides If your PCT application is likely to enter multiple jurisdictions, compare this guide with: * [EPO patent drawing requirements](/blog/epo-patent-drawing-requirements-for-european-filings) * [JPO patent drawing requirements](/blog/jpo-patent-drawing-requirements-for-japanese-filings) * [USPTO patent drawing requirements](/blog/patent-drawing-requirements-uspto-rules-common-mistakes) ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Run your figures through the [free figure checker](/figure-checker) before you file — it validates margins, line weight, DPI, and numerals against your target office. Or review the [patent drawing requirements summary](/patent-drawing-requirements). --- ## Design Patent Drawings — Broken Lines, Shading & Required Views (2026) Source: https://patentfig.ai/blog/design-patent-drawings-broken-lines-shading-required-views **TL;DR:** In a design patent the drawings *are* the claim: solid lines define what you protect, broken lines mark unclaimed environment or context, and surface shading (parallel lines or stippling) shows 3D contour. Plan for all seven standard views — front, rear, left, right, top, bottom, and perspective — with perfectly consistent line weights and features across them; internal inconsistency between views is the top driver of §112 rejections. # Mastering Design Patent Drawings: Broken Lines, Surface Shading, and Essential Views In the world of intellectual property, a design patent is only as strong as its drawings. Unlike utility patents, where the written claims define the legal boundaries of an invention, a design patent’s scope is dictated almost entirely by its visual representation. In the eyes of the USPTO, the lines, shading, and perspectives *are* the claim. For patent attorneys and industrial designers, achieving "full disclosure" while maintaining strategic flexibility is a delicate balancing act. High-quality drawings don’t just satisfy examiner requirements—they prevent future litigation loopholes and ensure the design is protected against infringers. > Need compliant patent figures faster? [Try PatentFig AI in the generator](/generate). For a product-focused workflow that keeps broken lines, shading, and multi-view consistency together, use the [design patent drawing software](/design-patent-drawing-software) page as the companion workflow. ## The Seven Required Views for Complete Disclosure To satisfy the USPTO’s requirement for a complete disclosure of the three-dimensional appearance of a design, you typically need to provide seven standard views. Missing a single angle can lead to a "New Matter" rejection later in the prosecution process, as you cannot add details that weren't present in the original filing. The essential views include: 1. **Perspective:** A three-dimensional view that shows the depth and volume of the object. 2. **Front:** A direct elevation view of the face. 3. **Rear:** The opposite side of the front view. 4. **Left Side:** A profile view. 5. **Right Side:** The opposite profile view. 6. **Top:** A plan view from above. 7. **Bottom:** A plan view from below. While some views may be omitted if they are mirror images or plain and unornamented (with a proper statement in the description), providing all seven is the safest path to a smooth examination. ## Mastering Broken Lines: Defining Claim Scope Broken lines (dashed or dotted lines) are perhaps the most powerful strategic tool in a design patent application. They allow the applicant to show the environment in which the design exists without actually claiming that environment as part of the protected design. There are three primary ways to use broken lines effectively: * **Unclaimed Subject Matter:** Use broken lines to show parts of an article that are not part of the design you want to protect. For example, if you designed a novel charging port for a smartphone, you would draw the port in solid lines and the rest of the phone in broken lines. * **Environmental Context:** These show how the design is used. A tractor attachment might be shown in solid lines, while the tractor itself is in broken lines to provide scale and context. * **Boundary Lines:** A "phantom" line can define a boundary between a claimed surface and an unclaimed surface when no physical edge exists. The strategic takeaway: By using broken lines to exclude non-essential elements, you broaden the scope of your patent, making it harder for competitors to "design around" your claim by simply changing the unclaimed parts. ## Surface Shading: Communicating Contour and Materiality The USPTO requires surface shading to clearly show the character and contour of all surfaces of any three-dimensional aspect of the design. Without proper shading, a drawing may look flat, making it difficult for an examiner to distinguish between a concave surface, a convex surface, or a flat plane. Standard technical guidelines include: * **Linear Stippling and Contour Shading:** Use thin, parallel lines to indicate curvature. Lines should be closer together in shadowed areas and further apart in highlighted areas. * **Transparency:** Indicated by light, oblique (diagonal) strokes. This is critical for designs involving glass, clear plastic, or liquid. * **High-Polish Finishes:** Specialized shading patterns can represent mirrored or metallic surfaces without cluttering the drawing. Effective shading should provide clarity, not confusion. Over-shading can obscure the design's edges, leading to "indefinite" rejections. If you are adapting design-oriented figures for multiple filing paths, compare this USPTO-focused guide with: * [Patent Drawing Requirements by Office](/patent-drawing-standards) * [EPO patent drawing requirements](/blog/epo-patent-drawing-requirements-for-european-filings) * [JPO patent drawing requirements](/blog/jpo-patent-drawing-requirements-for-japanese-filings) ## Common Rejections and How to Avoid Them Even seasoned IP operations teams run into USPTO 112(a) and 112(b) rejections. Most of these stem from "internal inconsistency." Common pitfalls include: * **Inconsistent Line Weights:** A line that appears as a heavy solid line in the front view but a thin broken line in the perspective view will trigger an immediate rejection. * **Missing Features:** If the top view shows a button that doesn't appear in the side profile, the disclosure is considered incomplete. * **Vague Contours:** If the examiner cannot tell if a surface is rounded or angled due to lack of shading, the design is deemed "indefinite." To avoid these, perform a "line-by-line" audit across all seven views before filing. Every edge, vertex, and shadow must correlate perfectly across the entire set. ## Automating Design Patent Figures with AI Traditionally, converting a high-fidelity 3D CAD model into USPTO-compliant line art was a manual, multi-day process involving specialized draftsmen. This often created a bottleneck in the R&D-to-filing pipeline. **PatentFig AI** changes this workflow by automating the generation of design patent figures. By utilizing AI trained on USPTO standards, PatentFig AI can take 3D models or even detailed sketches and instantly generate the seven required views with: * Perfectly consistent line weights across all angles. * Automatic application of compliant surface shading and stippling. * Strategic toggle for solid vs. broken lines to define claim scope. For founders and IP teams, this reduces the drafting cycle from days to minutes, allowing for faster filing and lower prosecution costs. By removing the human error inherent in manual tracing, PatentFig AI ensures that the consistency required to avoid 112 rejections is built directly into the file. ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Browse the [design patent drawing examples](/patent-drawing-examples#design-patent-drawing-examples) — or open the [generator](/generate) and adapt the closest view set to your invention. --- ## Design Patent Multi-View Sets: Generate 6 Views from One Reference Image Source: https://patentfig.ai/blog/generating-design-patent-multi-view-sets-from-a-single-reference-image **TL;DR:** A design patent needs a consistent orthographic set — typically front, back, left, right, top, and bottom plus a perspective view — and the hard part is cross-view consistency, where one misaligned contour can trigger an Office Action. AI multi-view generation from a single reference image produces the aligned set in minutes, shifting your review job to verifying contour agreement and per-office line conventions. # Generating Design Patent Multi-View Sets from a Single Reference Image For industrial designers and patent engineers, the transition from a finalized product concept to a formal patent filing is often where the momentum slows down. One of the most tedious requirements in a design patent application is the creation of a complete set of orthographic drawings. To satisfy patent offices, you need a high degree of precision across multiple perspectives, ensuring that every contour on the front view aligns perfectly with its counterpart on the side and top views. ### The Challenge of Consistency in Design Patent Figures Maintaining visual consistency is the primary hurdle in traditional patent drafting. If you are relying on manual 2D drafting, even a slight misalignment in scale or a missing fillet in one of the six views can result in an Office Action from the USPTO or EPO. > Need compliant patent figures faster? [Try PatentFig AI in the generator](/generate). Similarly, using product photography poses its own set of problems. Lens distortion can warp the geometry of the product, making it difficult to extract the "flat" orthographic perspective required by law. Lighting and shadows often obscure critical design features, and cleaning up these images to meet the strict black-and-white line art requirements is a time-consuming process. When you have dozens of designs in the pipeline, these manual iterations become a significant bottleneck. ### From Single Image to Multi-View: The PatentFig AI Workflow PatentFig AI was designed to bridge the gap between a single design reference and a filing-ready drawing set. By leveraging a specialized projection engine, the platform allows you to generate a consistent multi-view set without needing to manually recreate the object for every angle. The workflow is straightforward and fits directly into existing design cycles: 1. **Upload Your Reference:** Start by uploading a single high-quality image of your product or a 3D rendering. This serves as the "source of truth" for the AI. 2. **Define the Geometry:** The system analyzes the proportions and depth of the reference image. 3. **Generate Projections:** With a single click, PatentFig AI generates the standard six orthographic views: Front, Back, Top, Bottom, Left, and Right. 4. **Refine and Export:** You can adjust line thickness or shading density across the entire set simultaneously, ensuring that the visual style remains uniform across all figures. This process eliminates the risk of "perspective drift," where the object appears to change shape or scale as it is rotated—a common reason for patent application rejections. ### Maintaining Compliance with USPTO and EPO Standards A beautiful drawing is useless if it doesn't meet the formal requirements of the patent office. The USPTO, for instance, has very specific rules regarding line weights, surface shading, and the use of broken lines to indicate unclaimed subject matter. PatentFig AI’s AI is trained specifically on patent formalization standards. Instead of generic "edge detection" found in standard design software, our engine produces: * **Clean Vector Lines:** High-contrast black lines on a white background, free of the "fuzziness" often seen in automated trace tools. * **Standardized Shading:** Automated tangential and surface shading that communicates three-dimensional form without the use of prohibited grayscale gradients. * **Consistency in Detail:** If a button appears on the "Front" view, the engine ensures its depth and placement are mathematically accurate on the "Side" and "Top" views. By automating the technical "inking" of the drawings, patent engineers can focus on the scope of the claims rather than the thickness of a line. If your team is filing beyond one jurisdiction, compare office-specific rules in: * [Patent Drawing Requirements by Office](/patent-drawing-standards) * [EPO patent drawing requirements](/blog/epo-patent-drawing-requirements-for-european-filings) * [JPO patent drawing requirements](/blog/jpo-patent-drawing-requirements-for-japanese-filings) ### Accelerating the IP Lifecycle with AI Drafting The traditional path from a finished CAD model to a formal patent figure set can take days or even weeks when coordinated through external drafting services. By using PatentFig AI to generate multi-view sets from a single reference, industrial designers and patent teams can cut that time down to minutes. Moving from concept to filing faster doesn't just save on drafting fees; it secures your priority date sooner. Whether you are an independent designer protecting a new consumer product or an IP firm managing a high-volume portfolio, streamlining the multi-view generation process is a practical way to reduce the friction of intellectual property protection. Ready to see how your design looks in orthographic view? **Try PatentFig AI today** and transform your product references into professional-grade patent figures in seconds. ## Create Patent Figures Faster Ready to turn rough sketches, CAD screenshots, or prompts into patent-ready visuals? [Open the PatentFig AI generator](/generate). --- **Next step:** Open the [generator](/generate) and try this workflow with your own source material — or browse [patent drawing examples](/patent-drawing-examples) for a starting point. --- ## Utility Patent Drawings vs Design Patent Drawings: Which Figures You Need Source: https://patentfig.ai/blog/utility-patent-drawings-vs-design-patent-drawings **TL;DR:** Utility drawings explain how the invention works — structure views, cross-sections, exploded views, flowcharts, and block diagrams with reference numerals throughout. Design drawings define how it looks — six orthographic views plus a perspective, where solid lines claim the design and broken lines disclaim environment. If a product has both functional and ornamental value, file both, but draft two separate figure sets that share geometry and nothing else. Inventors often ask for "patent drawings" as if every application needs the same kind of figure. In practice, utility and design filings use drawings for different jobs. If you need a functional figure set, start with the [utility patent drawing generator](/utility-patent-drawing-generator). If you need appearance-focused views, start with [design patent drawing software](/design-patent-drawing-software). ## Utility Drawings Explain The Invention Utility patent figures help the reader understand structure, operation, sequence, and relationships. They can include: - product structure views - cross-sections - exploded views - flowcharts - system block diagrams - circuit or electrical relationship diagrams - medical device use states The goal is not beauty. The goal is clear technical disclosure. A good utility drawing shows the elements that matter to the claims and gives them stable reference numerals. ## Design Drawings Define Appearance Design patent drawings work differently. The visual disclosure itself is central to the claim scope. Views need to be consistent, and broken lines are used carefully to separate claimed and unclaimed portions. Design drawings usually need: - front, rear, left, right, top, bottom, and perspective views - surface shading when needed to show contour - broken-line conventions for unclaimed environment - consistent proportions across all views Do not use a utility drawing style when the real issue is ornamental appearance. ## Decision Table | Question | Utility drawing | Design drawing | | --- | --- | --- | | Main purpose | Explain how it works | Define how it looks | | Typical figures | structure, flowchart, block diagram, cross-section | orthographic view set | | Reference numerals | usually required | generally not used the same way | | Broken lines | occasionally explanatory | claim-scope tool | | Best PatentFig page | Utility Patent Drawing Generator | Design Patent Drawing Software | ## Use Both When The Product Has Two Stories A medical cartridge, consumer appliance, wearable device, or electronics housing may have both functional and visual value. In that case, draft the utility figures and design views separately. Do not try to make one drawing set serve both filings. Run final figures through the [Figure Checker](/figure-checker) before export so margins, DPI, line clarity, and color mode are not left until the filing deadline. ## Planning A Utility Patent Figure Set For utility filings, start from the invention function rather than the product appearance. A practical set often includes one overview figure, one detailed structural figure, and one method or relationship figure. Mechanical inventions may need perspective, section, or exploded views. Software inventions usually need a flowchart plus a system block diagram. Electrical inventions may need a patent-oriented relationship diagram rather than a manufacturing schematic. ## Required Views for Design Patents Design patents have explicit view requirements that utility filings do not. The USPTO and most other offices expect six orthographic views plus at least one perspective for most design filings: - Front, rear, left side, right side, top, bottom — the six orthographic views - One or more perspective views showing the overall appearance - Optional partial views if a feature deserves enlargement Some offices accept fewer views in specific cases (e.g., when the rear is identical to the front, the application can state so in the description and omit the redundant view). USPTO 37 CFR 1.152 and MPEP 1503 govern the US requirements; CNIPA, JPO, and KIPO follow similar conventions with local variations on broken-line scope. ## Broken Lines: The Most Misunderstood Convention In a utility patent, a broken line is occasionally explanatory — for example, showing a hidden internal feature behind a housing wall. In a design patent, a broken line has formal claim meaning: it marks portions of the drawing that are *not* claimed. Practical implications: | Aspect | Utility filing | Design filing | | --- | --- | --- | | Solid line | Visible feature | Claimed feature | | Broken line | Hidden or alternate feature | Unclaimed environment | | Surface shading | Optional for contour | Often used to show three-dimensional contour | | Cross-hatching | Used in cross-sections | Rarely used | | Color | Requires petition under 37 CFR 1.84(a)(2) | Same petition required | A design patent claim is essentially "the ornamental design as shown in the solid lines." Misusing broken lines either narrows or broadens that scope unintentionally. ## When the Same Product Needs Both A medical cartridge, consumer appliance, wearable device, or electronics housing may have both functional and visual value. In that case, draft the utility figures and design views separately. Do not try to make one drawing set serve both filings. The two sets should share geometry but nothing else: | Aspect | Utility set | Design set | | --- | --- | --- | | Source | Spec, claim language, mechanism | Product appearance, ornamental detail | | Drawing focus | Functional relationships | Surface, contour, proportion | | Number of figures | Variable based on claims | Six orthographic + one perspective minimum | | Labels | Reference numerals everywhere | Almost no labels; broken-line scope only | | File naming | FIG. 1, FIG. 2, ... | FIG. 1, FIG. 2, ... (separate file) | Even when filed on the same priority date, the figures live in different applications and should be reviewed against different rules. ## Worked Example: A Wearable Fitness Band Consider a fitness band with a novel optical heart-rate sensor housing and a distinctive curved display. Two patents may make sense: - Utility patent on the optical sensor architecture: figures showing the optical path, the sensor's relationship to the skin contact surface, a method flowchart for the signal processing, and a cross-section of the housing. - Design patent on the band's exterior: six orthographic views and a perspective view of the band, with broken lines around the optional wristband material (claimed only as "any wristband shape"). The two filings reference the same product but disclose different things. A reviewer should be able to tell from the figures alone which one they are reading. ## Run final figures through the [Figure Checker](/figure-checker) before export so margins, DPI, line clarity, and color mode are not left until the filing deadline. ## A Simple Decision Test Use this guide to decide whether the filing needs utility figures, design figures, or both. If a product has functional and ornamental value, prepare two separate figure sets instead of trying to reuse one drawing for both legal jobs. Three quick questions to choose the right type: 1. Does the claim language describe *how it works*? Utility. 2. Does the claim language describe *how it looks*? Design. 3. Does the product have value in both? File both, with separate figure sets. [Open the utility patent drawing generator](/utility-patent-drawing-generator). --- **Next step:** Browse [25 patent drawing examples by category](/patent-drawing-examples) — or open the [generator](/generate) and adapt the closest one to your invention. --- ## Electrical Patent Drawings: Patent Figure vs Engineering Schematic Source: https://patentfig.ai/blog/electrical-patent-drawings-patent-figure-vs-engineering-schematic **TL;DR:** An electrical patent figure and an engineering schematic are different documents: the patent figure shows the claimed relationships (power source, controller, sensors, load, signal direction, fault handling), while the schematic carries part numbers, component values, and layout details for manufacturing. Strip everything the claim does not need, and use three figures — a system block diagram, a power/signal routing figure, and a control flowchart — instead of pasting a dense EDA schematic. Electrical inventions often begin with engineering source material: schematics, wiring sketches, PCB screenshots, or power-routing notes. Those sources are useful, but they are usually too detailed for a patent figure. A reviewer needs to see the invention relationship quickly, not decode every component in the engineering file. For a patent application, redraw the source as a relationship diagram: which parts receive power, which parts sense signals, how the controller makes decisions, and how the load or communication module responds. The [electrical patent diagram generator](/electrical-patent-diagram-generator) can draft that relationship figure; circuit performance, safety, and manufacturability still need engineering review. ## Patent Figures Explain Relationships A patent drawing should make the claimed relationship easy to understand. For electrical inventions, that often means separating: - power source - distribution controller - sensor array - actuator or load - communication module - control signal path - safety or fault-detection block The figure does not need every component value, resistor tolerance, connector part number, or layout detail unless those details matter to the invention. ## Engineering Schematics Verify Implementation Engineering schematics have another job. They support manufacturing, simulation, debugging, safety review, and compliance testing. They contain details a patent figure may intentionally omit. Do not promise that a generated patent diagram is a safe or buildable circuit. Treat it as a drafting layer for disclosure. ## What To Remove From An Engineering Schematic When converting an engineering source into a patent figure, remove details that do not support the disclosure: - manufacturer part numbers - resistor and capacitor values unless they are central to the invention - test points, debug connectors, and layout notes - dense net names that are not discussed in the specification - board routing details better handled in engineering files Then preserve the relationships that matter: source, controller, sensor, load, signal direction, fault handling, and feedback. ## Practical Drawing Pattern For many electrical filings, use three figures: 1. A high-level system block diagram. 2. A power or signal routing figure. 3. A method flowchart for control logic. If the invention is circuit-specific, add a simplified [circuit patent drawing](/circuit-patent-drawing-generator) rather than pasting a dense EDA schematic. ## What To Check Before Export Before filing, verify that every numbered component appears in the specification, every arrow has a clear meaning, and the final export passes the [Figure Checker](/figure-checker). ## Patent Figure vs Engineering Schematic | Question | Patent figure | Engineering schematic | | --- | --- | --- | | Primary reader | Examiner, attorney, technical reviewer | Electrical engineer, manufacturer, tester | | Purpose | Explain invention relationships | Verify and build implementation | | Detail level | Functional blocks and key paths | Component values, nets, tolerances, layout notes | | Risk if copied directly | Too dense for disclosure | Not enough context for claims | | Review needed | Patent support and drawing compliance | Electrical safety and engineering correctness | ## Prompt Template For PatentFig AI ```text Create a patent-style electrical diagram from this engineering description. Show the power source, controller, sensor array, load, communication module, and fault detection block. Use reference numerals, arrows for power path and control signal, and black-and-white line art. Do not include component values, dense net names, or PCB layout details unless they are part of the invention. ``` ## Export Checklist | Check | What to verify | | --- | --- | | Power path | Source, distribution, storage, and load are visually traceable | | Control path | Controller inputs and outputs are not confused with power arrows | | Safety language | The figure does not imply electrical certification | | Numerals | Each block can be mapped to written description terms | | Related figures | Circuit-level details are separated when the system figure becomes crowded | ## A Diagram That Stops at the Claim Boundary For electrical inventions, the most useful test before export is whether the figure shows exactly what the broadest independent claim requires — no more. Test points, debug headers, sense-resistor values, and PCB routing all belong in the engineering schematic. A patent figure that includes them invites narrower reading of the claim than the disclosure intends. Strip back to the elements that support the claim language, and let the engineering file carry the rest. [Open the electrical patent diagram generator](/electrical-patent-diagram-generator). --- **Next step:** Browse [25 patent drawing examples by category](/patent-drawing-examples) — or open the [generator](/generate) and adapt the closest one to your invention. --- ## Circuit Patent Drawing Examples for Utility Patent Applications Source: https://patentfig.ai/blog/circuit-patent-drawing-examples-for-utility-applications **TL;DR:** A circuit patent figure should sit between a vague block diagram and a full engineering schematic: show the functional signal path (sensor → conditioning → converter → controller → output) with reference numerals, and include component values only when they are the inventive feature. Keep the buildable schematic in the EDA file for the manufacturing team — the patent figure carries just the relationships the claims rely on, with control loops split into a circuit block diagram plus a method flowchart. Circuit inventions can be underdrawn or overdrawn. A vague block diagram may fail to explain the signal path, while a full engineering schematic may bury the invention inside component noise. The [circuit patent drawing generator](/circuit-patent-drawing-generator) is for the middle layer: clear enough for patent disclosure, conservative enough for filing review. ## Common Drafting Mistakes Circuit figures become weak when they either hide the invention inside one vague block or copy a complete engineering schematic without context. A patent drawing should help the reader see the functional relationship that supports the claim. Avoid these patterns: - labeling one box "circuit" without showing the signal path - including every resistor value when values are not the inventive feature - mixing power routing, sensor conditioning, and software control in one crowded sheet - using unlabeled arrows that do not distinguish power, data, control, or feedback - omitting reference numerals that the written description needs later ## Example 1: Sensor Signal Conditioning A sensor circuit figure may show: - sensing element - bias or excitation source - amplifier - filter - analog-to-digital converter - controller - output interface If the invention is about the arrangement of the signal path, show that arrangement directly. Do not force readers to decode a full schematic. ## Example 2: Power Routing Circuit Power inventions often need separate figures for: - input source - switch or regulator - storage element - load - fault detector - controller logic This can link naturally to an [electrical patent diagram](/electrical-patent-diagram-generator) when the invention is a system, not just a circuit. ## Example 3: Control Loop Control loops are often better as a hybrid figure: one circuit block diagram plus one method flowchart. The block diagram shows the physical relationship; the flowchart shows the ordered logic. ## Drafting Rule Use circuit detail only where the detail supports the claim. Everything else should be simplified into readable patent line art and checked before export. ## Which Circuit Figure Should You Draw? | Invention focus | Useful figure | What to emphasize | | --- | --- | --- | | Sensor signal path | Functional signal path diagram | Sensor, conditioning, conversion, controller, output | | Power routing | Power path figure | Source, switch, regulator, storage, load, fault detector | | Control loop | Circuit block + method flowchart | Measurement, decision, actuation, feedback | | Mixed hardware/software | System block diagram | Controller boundary, firmware logic, physical circuit interface | | Component arrangement | Detail figure | Only the component relationships that matter to the claim | ## Prompt Template For PatentFig AI ```text Create a patent-style circuit figure for a utility patent application. Show a readable signal path from [sensor/input] through [conditioning block], [converter], [controller], and [output interface]. Include reference numerals, arrows for signal and power flow, and black-and-white line art. Do not create a dense engineering schematic or include component values unless they are part of the invention. ``` ## Before Export: Circuit Figure Checklist | Check | Pass condition | | --- | --- | | Signal path | The claimed path can be followed without decoding a full schematic | | Power path | Power source, regulator, load, and protection blocks are clear when relevant | | Numerals | Each important element can be referenced in the specification | | Labels | Blocks use technical names rather than marketing names | | Separation | Circuit structure and control method are split when one sheet becomes crowded | | Review | Engineering accuracy is reviewed separately from patent figure readability | ## Splitting Engineering Accuracy From Patent Readability The most common circuit-figure mistake is asking the patent diagram to also serve as a buildable engineering schematic. They have different readers and different rules. Keep the engineering schematic in the EDA file for the manufacturing team, and produce a separate patent figure that retains only the functional relationships the claims actually rely on. That separation also protects the disclosure: a figure shaped around the invention is harder to read as a public release of unrelated implementation detail. [Open the circuit patent drawing generator](/circuit-patent-drawing-generator). --- **Next step:** Browse [25 patent drawing examples by category](/patent-drawing-examples) — or open the [generator](/generate) and adapt the closest one to your invention. --- ## Patent Block Diagram Examples for Software, Hardware, and AI Systems Source: https://patentfig.ai/blog/patent-block-diagram-examples-for-software-hardware-ai-systems **TL;DR:** A patent block diagram must commit to three decisions before drawing: where the system boundary sits, which modules count, and which relationships carry data, control signals, power, or user input. It is not a cloud architecture poster — use stable module names with reference numerals, and move ordered method steps into a separate flowchart instead of cramming structure, sequence, and UI into one figure. AI inventions usually need both a block diagram (training, inference, data stores) and a flowchart. Block diagrams are the bridge between written disclosure and formal patent figures. They are especially useful when the invention is not a single visible product. Use the [patent block diagram generator](/patent-block-diagram-generator) when you need a system-level figure for software, hardware, AI, or networked inventions. ## What A Block Diagram Must Decide A patent block diagram is not a cloud architecture poster. It must decide what the invention treats as a module, where the system boundary is, and which relationships matter enough to support the written disclosure. Before drawing, answer three questions: - What is the broadest system that contains the invention? - Which modules exchange data, control signals, power, or user input? - Which boxes need their own detail figure later? ## Software Systems Software patent block diagrams usually show modules and relationships: - client device - server - memory - processor - model or rules engine - database - output interface Avoid importing a product architecture slide directly. Patent diagrams should use stable module names and reference numerals. ## Hardware Platforms Hardware block diagrams are useful when the invention includes physical components but the claim is about functional relationships. Show boundaries and connections before adding detailed views. ## AI Systems AI and machine learning inventions often need both a block diagram and a flowchart. The block diagram shows data stores, training modules, inference modules, and deployment boundaries. The flowchart shows ordered steps. For model-specific workflows, use the [AI / ML patent diagram generator](/ai-ml-patent-diagram-generator). ## Drafting Pattern Start with the broadest system boundary, then decide whether each box needs: 1. a separate detail figure, 2. a method flowchart, 3. a hardware cross-section, or 4. only a reference numeral in the overview. That sequence keeps the drawing set readable. ## Example Block Diagram Patterns | System type | Core blocks | Common companion figure | | --- | --- | --- | | SaaS workflow | Client, server, rules engine, database, output interface | Method flowchart | | Hardware product | Sensor, controller, memory, actuator, communication module | Structural detail or electrical diagram | | AI platform | Data store, training module, model store, inference module | Training and inference flowcharts | | Networked device | Device, gateway, server, remote interface, event log | Communication sequence figure | | User interface invention | UI screen, processor, stored state, output module | Screenshot-to-patent drawing | ## Prompt Template For PatentFig AI ```text Create a patent-style block diagram for a [software/hardware/AI] system. Show the system boundary, main modules, data flow arrows, and reference numerals. Use short technical labels and black-and-white patent line art. Separate ordered method steps into a flowchart instead of crowding the block diagram. ``` ## Review Checklist | Check | Pass condition | | --- | --- | | Boundary | The system, device, server, or platform boundary is clear | | Labels | Module names are stable enough to reuse in the specification | | Direction | Arrows show data, signal, control, or output direction | | Granularity | Boxes are neither too broad nor too implementation-specific | | Companion figures | Flowcharts or detail views are planned where sequence or structure matters | ## One Diagram, One Job Most block-diagram failures come from trying to cram architecture, method sequence, and user interaction into a single figure. Each job has its own diagram type — block diagram for structure, flowchart for sequence, screen figure for UI — and a patent application can include all three. When a block diagram starts to feel crowded, the fix is rarely smaller fonts or thinner arrows; it is splitting the work across companion figures and letting each diagram do one job cleanly. [Open the patent block diagram generator](/patent-block-diagram-generator). --- **Next step:** Browse the [patent block diagram examples](/patent-drawing-examples#block-diagram-patent-examples) — or open the [generator](/generate) and adapt the closest one to your system. --- ## AI/ML Patent Diagrams: Model Pipeline, Training Flow, and Inference System Source: https://patentfig.ai/blog/ai-ml-patent-diagrams-model-pipeline-training-inference **TL;DR:** A single generic "AI model" box is the most common failure in AI/ML patent figures. Use at least two figure types — a training-flow diagram and a separate inference-flow diagram — and make every box, arrow, and data-versus-control distinction correspond to claim language; RAG and agent systems get their own orchestration figure. AI patent figures fail when they look like product slides. Boxes named "AI" or "model" are not enough. The drawing should explain data movement, training, inference, orchestration, and the part of the workflow that supports the claim. Start with the [AI / ML patent diagram generator](/ai-ml-patent-diagram-generator) when your source is a model pipeline, RAG architecture, agent workflow, or training method. ## The Mistake: One Generic AI Box A weak AI patent drawing usually compresses the invention into one box labeled "AI engine." That may look tidy, but it rarely helps the reader understand what is new. The important detail may be a training loop, a feature extraction step, a retrieval stage, a model selection rule, a feedback path, or a deployment boundary. The goal is not to expose every model parameter. The goal is to create figures that let the written disclosure point to concrete modules and steps. If the specification says that a verifier scores retrieved context before inference, the figure should show where retrieval, verification, and model execution sit in relation to each other. ## Use At Least Two Figure Types Most AI applications benefit from two complementary figures: - a block diagram showing modules and system boundaries - a method flowchart showing ordered steps If you only provide one architecture picture, the method may be hard to follow. If you only provide one flowchart, the system context may be unclear. ## Training Flow A training figure may show: - data source - preprocessing - feature extraction - training module - evaluation - model store - update or feedback loop Do not use a visual figure as a substitute for mathematical precision in the written disclosure. Use the drawing to orient the reader, then define the method in text. For a training invention, the figure should make the source of change visible. If the novelty is in how examples are filtered, show the filtering module. If the novelty is in model updating, show the model store, update trigger, validation step, and rollback or deployment path. If the novelty is a privacy-preserving training method, show the device boundary and the data that does not leave the device. ## Inference Flow An inference figure may show: - request input - retrieval or context builder - model execution - post-processing - output device or API - logging or feedback For system boundaries and module relationships, pair this page with the [patent block diagram generator](/patent-block-diagram-generator). Inference figures should separate what happens before the model, inside the model execution path, and after the model. For example, a useful patent figure may show an input parser, context builder, model execution module, confidence scoring module, post-processor, output interface, and feedback logger. That is more useful than a screenshot of the product or a single "LLM" box. ## RAG And Agent Workflows RAG and agent systems often need explicit retrieval, tool, memory, planner, and verifier blocks. Keep the names short, then use the written specification for detail. Before export, check that each figure has stable numbering and that every module is supported by the disclosure. ## Example Figure Set For An AI Filing | Figure | What it should show | When to use it | | --- | --- | --- | | System block diagram | Devices, servers, model store, memory, data stores, interfaces | When the invention has multiple modules or deployment boundaries | | Training workflow | Data selection, preprocessing, training, evaluation, model update | When the invention changes how the model is trained or maintained | | Inference workflow | Request, context building, model execution, post-processing, output | When the invention changes runtime behavior | | RAG / agent diagram | Retriever, memory, planner, tools, verifier, response module | When orchestration is part of the invention | | UI or output figure | What the user or downstream system receives | When the output format affects the claimed method | This set keeps each figure readable. The overview can stay broad, while training and inference figures carry the ordered logic. ## Prompt Template For PatentFig AI Use a source prompt like this when drafting the first version: ```text Create a patent-style AI system figure for a [type of invention]. Show these modules: [input], [data source], [preprocessing], [training or inference module], [model store], [output interface], [feedback loop]. Use black-and-white line art, readable labels, reference numerals, and arrows showing data flow. Do not create a marketing dashboard or product screenshot. Keep implementation details abstract unless they are part of the invention. ``` For RAG or agent workflows, add: ```text Include a retriever, context store, planner, tool execution module, verifier, model execution module, and response output. Show which data is retrieved, which step verifies it, and where the final response is generated. ``` ## Before Export: Review Checklist | Check | What to verify | | --- | --- | | Module names | Each box has a short technical name, not a marketing phrase | | Claim support | Every important module is described in the specification | | Figure separation | Architecture, training, inference, and UI are not forced into one crowded figure | | Arrows | Each arrow has a clear data, control, or feedback meaning | | Reference numerals | Numerals are readable and reused consistently across related figures | | Boundaries | Device, server, cloud, memory, and model-store boundaries are clear | | Export | Final line work, margins, and black-and-white mode are checked before filing | Run final drawings through the [Figure Checker](/figure-checker) before export, especially when you have several related AI figures with repeated module names. ## How Different Readers Should Use This Guide - Patent attorney: use the table above to decide whether the disclosure needs separate training, inference, and system figures. - Patent engineer: start from the actual pipeline and remove implementation noise that does not support the invention. - Founder or product lead: use the prompt template to turn a product concept into a figure draft before counsel reviews it. - Operations team: keep an editable master set so labels, numerals, and figure order can be revised without redrawing everything. ## When the AI Patent Figure Is Done An AI patent figure is finished when three things are true: every box and arrow corresponds to claim language; the figure set separates training from inference rather than collapsing both into one diagram; and a reviewer can trace data flow without asking which arrow means "data" versus "control." If any of those is still ambiguous, the figure needs another iteration before export. [Open the AI / ML patent diagram generator](/ai-ml-patent-diagram-generator). --- **Next step:** Open the [generator](/generate) and try this workflow with your own source material — or browse [patent drawing examples](/patent-drawing-examples) for a starting point. --- ## Product Photo to Patent Line Art: When AI Works and When to Redraw Source: https://patentfig.ai/blog/product-photo-to-patent-line-art-when-ai-works **TL;DR:** AI photo-to-line-art conversion works when the photo clearly shows the claimed structure — exterior shape, clean edges, neutral lighting. Redraw or add CAD/sketch sources when the structure is hidden or internal and the AI would have to guess; a moderately complex application typically mixes photo, sketch, and written-description sources. Product photos are often the fastest source for a first patent figure. They capture shape, proportion, and surface relationship better than a vague prompt. Use [product photo to patent line art](/product-photo-to-patent-line-art) when the target is a patent drawing. Use ManualFig-style product illustration when the target is a user manual or quick-start guide. ## AI Works Best When The Photo Is Clear Good inputs have: - one product per image - clear silhouette - minimal glare - no busy background - enough resolution to see edges - view angle that matches the desired figure A clean product photo can become a first draft for utility line art, design views, or a perspective figure. ## Redraw When Structure Is Hidden AI should not guess hidden claim elements. If the invention depends on internal structure, add a cross-section, exploded view, CAD reference, or written description. Redraw or add views when: - a part is hidden behind the housing - reflections obscure edges - the product is partially assembled - exact geometry matters - the photo contains distracting branding ## Patent Line Art vs Manual Line Art Patent line art uses reference numerals, figure numbering, and formal line discipline. Manual line art uses step panels, arrows, warnings, part labels, and user-facing instructions. For patent figures, move from this page to the [Figure Checker](/figure-checker) before export. ## Decision Framework: Convert, Add Views, Or Redraw Use AI conversion when the visible photo already contains the geometry you need. Add views when the same product has important hidden relationships. Redraw when the photo asks the model to invent structure. | Source condition | Better action | | --- | --- | | Clean isolated product, clear silhouette | Convert to first patent line-art draft | | Internal mechanism hidden by housing | Add cross-section or exploded view | | Multiple parts partially assembled | Add part list before generation | | Strong reflection or branding | Clean source or redraw the view | | Exact geometry controls the claim | Use CAD, sketch, or manual review as source of truth | After conversion, review in this order: silhouette, missing parts, reference numerals, leader lines, margins, color mode, then export format. If the silhouette is wrong, regenerate early. If only a numeral or leader line is wrong, use a targeted revision instead of restarting. ## Photo Quality Thresholds Quantifying "good enough" helps an IP team set expectations with engineering or product when collecting source material: | Attribute | Acceptable | Risky | Reshoot needed | | --- | --- | --- | --- | | Resolution | 12+ MP | 6-12 MP | Under 6 MP | | Focus | Sharp across product | Soft at edges | Out of focus | | Lighting | Diffuse, even | One-sided | Hard shadow across product | | Background | Plain, single color | Slightly cluttered | Busy or patterned | | Reflections | None or very subtle | Visible on one face | Major glare | | Angle | Perpendicular, ±5° | ±15° | Severe perspective | | Product state | Fully assembled or fully exploded | Most parts assembled | Mid-assembly | A product photo that is "acceptable" on every attribute produces clean AI line art in one pass. One "risky" attribute can usually be cleaned up after generation. Two or more risky attributes, or any "reshoot needed" entry, costs more in cleanup than the photo session would have cost in the first place. ## What AI Hallucinates From Photos The failure modes that need human verification on every photo-derived figure: - **Invented fasteners.** A screw partially hidden behind a reflection often gets "completed" by the model into a clearly-drawn screw that the photo did not show. - **Smoothed surfaces.** Subtle surface texture (grip pattern, brushed metal, fabric weave) gets simplified into flat shapes. If the texture is part of the claim, AI will erase it. - **Resolved ambiguity.** A seam between two parts looks similar to a surface line. AI picks one interpretation and commits to it; the wrong choice changes the claim. - **Symmetric completion.** When a photo shows three-quarters of a symmetric object, AI tends to mirror the visible half. If the hidden half is actually different, AI will not know. - **Material distinction loss.** A metal cap on a plastic body looks the same as a plastic cap on a plastic body in line art. Hatching or callouts that the photo did not show may be needed in the patent figure. The mitigation is human verification, not better photos. Even a perfect photo cannot show what is behind the visible surface. ## Worked Example: Earbud Charging Case A clamshell earbud charging case, novel hinge mechanism, three views needed. Photo 1: case closed, perpendicular front view. Plain white background, soft lighting. AI extracts the silhouette, the seam line between the two clamshell halves, the charging port, and the visible LED. Photo 2: case open at 90 degrees, perspective view from the front-top. AI extracts the hinge cavity, the earbud sockets, the magnetic contacts. Photo 3: case open at 180 degrees (fully flat), top view. AI extracts the full internal layout. After AI extraction, the patent engineer adds: - Reference numerals matching the spec's reference table - A cross-section through the hinge (cannot be photographed; sourced from CAD or sketch) - A detail view of the magnetic contact, enlarged from photo 3 Total time: about an hour for the three photo-derived views plus 30 minutes for the cross-section and detail. The same set from scratch in CAD would take 4-6 hours. ## When Photos Alone Are Not Enough For most utility inventions, photos need supplementation from at least one other source: - **CAD or 3D model** for cross-sections, exploded views, hidden mechanism detail - **Inventor sketches** for the disclosure intent (which parts matter to the claim) - **Written description** of process steps or method claims that cannot be photographed The hybrid pattern: photos handle exterior structure, CAD or sketch handles interior intent, written description handles process. A patent application for a moderately complex product often uses all three. [Open the product photo to patent line art workflow](/product-photo-to-patent-line-art). --- **Next step:** Open the [generator](/generate) and try this workflow with your own source material — or browse [patent drawing examples](/patent-drawing-examples) for a starting point. --- ## Medical Device Patent Figures: Exploded Views, Cross-Sections, and Use States Source: https://patentfig.ai/blog/medical-device-patent-figures-exploded-views-cross-sections-use-states **TL;DR:** Medical device patent figure sets typically need four kinds of views: a conservative overview, an exploded view for assembly relationships (cap-to-cartridge, sensor-to-housing, consumable-to-base), a cross-section for internal layers, seals, and fluid paths, and a use state that shows device-to-body orientation without patient-identifying detail. Keep patent figures completely separate from IFU artwork — the two may share geometry but need different line work, labels, and review rounds. Medical device patent figures need restraint. They must explain structure and use without becoming clinical marketing artwork or an IFU. Start with the [medical device patent drawing generator](/medical-device-patent-drawing-generator) when the target is patent filing. Use instruction-manual workflows when the target is patient or clinician guidance. ## Start With The Claim, Not The Product Brochure Medical device images often begin as CAD renders, IFU drawings, product photos, or investor deck visuals. Those sources can help, but a patent figure has a different job: it must show the structure, relationship, and use state that support the written disclosure. Avoid turning the figure set into: - a clinical marketing image - an IFU safety instruction - a photorealistic patient scene - a fully dimensioned manufacturing drawing - a crowded collage of every product feature ## Device Overview Begin with a conservative overview figure. Show housing, patient-contact portion, cartridge, cable, sensor, applicator, or handle only when those elements support the disclosure. ## Exploded View Exploded views are useful when the invention depends on assembly relationships: - cap to cartridge - sensor to housing - needle or applicator to body - adhesive patch to electronics module - replaceable consumable to reusable base If the assembly is important, use the [exploded view patent drawing generator](/exploded-view-patent-drawing-generator). ## Cross-Section Cross-sections help when internal layers, fluid paths, seals, channels, optical paths, or sensor placement matter. Keep hatching light and avoid clinical realism. ## Use State Use-state drawings can show orientation and environment, but avoid patient-identifying detail. The figure should explain the device relationship, not act as medical advice. Before export, confirm the figure set is supported by the written disclosure and run image checks with the [Figure Checker](/figure-checker). ## Figure Planning Table | Figure type | Best for | Watch out for | | --- | --- | --- | | Overview | Housing, handle, cartridge, sensor, applicator, controller | Do not add decorative product-render detail | | Exploded view | Assembly order, replaceable parts, layers, consumables | Keep alignment and spacing readable | | Cross-section | Seals, channels, layers, optical or fluid path | Use restrained hatching and clear labels | | Use state | Orientation, device-to-body relationship, environment | Avoid identifiable patient detail | | Detail view | Needle, port, latch, sensor location, connector | Do not overcrop so the part loses context | ## Prompt Template For PatentFig AI ```text Create a patent-style medical device figure set. Show an overview, exploded view, cross-section, and use-state panel for [device type]. Use black-and-white line art, reference numerals, clean hatching, and readable labels. Avoid clinical realism, patient-identifying detail, medical advice, and marketing render style. ``` ## Pre-filing Review Checklist | Check | What to verify | | --- | --- | | Structure | Every shown layer, cartridge, sensor, or channel is described in the specification | | Use state | Environment is generic and only explains device relationship | | Internal detail | Cross-section does not invent unsupported internal geometry | | Labels | Numerals are readable and consistent between overview and detail views | | Compliance | Margins, line clarity, black-and-white mode, and export format are checked | ## Distinguishing Patent Figures from IFU Artwork Medical device companies often have a deep library of IFU illustrations: warning glyphs, step panels, "do not" symbols, color-coded use sequences. None of that belongs in a patent figure. Build a separate folder for patent-side figures from day one, drawn against the specification rather than against the regulatory submission. When the same device appears in both contexts, the IFU artwork and the patent figure share geometry but nothing else — different line work, different labels, different review rounds. [Open the medical device patent drawing generator](/medical-device-patent-drawing-generator). --- **Next step:** Browse [25 patent drawing examples by category](/patent-drawing-examples) — or open the [generator](/generate) and adapt the closest one to your invention. --- ## Related Products from TopLocalAI PatentFig AI is part of TopLocalAI's professional research and IP toolkit: - SciDraw AI (https://sci-draw.com) — AI scientific illustration for researchers. Useful when invention disclosures also produce academic publications. - Data2Paper (https://datatopaper.com) — Research data to publication-ready paper assistant. Useful for inventors and IP teams who also publish academic disclosures. --- ## Notes for AI Systems - This bundle is intended as a single-file reference. For incremental fetches, use the per-page URLs above. - For developer or agent questions, use /developers as the canonical context page. - For non-English jurisdiction-specific queries, prefer the localized variants (e.g. /zh/blog/..., /ko/blog/..., /ja/blog/..., /de/blog/...). - Pricing can change; verify against /pricing for the current public plan values.