PatentFig AI API
SkillDocs & knowledgeGenerate patent-office-compliant figures via the PatentFig AI API, patent line art from text (PNG or SVG), vectorize drawings to SVG/DXF/vector PDF, AI-upscale images, and convert to filing-ready TIFF/PDF/PNG. Use when the user asks for patent drawings, patent figures, invention diagrams, USPTO/EPO-compliant line art, vectorizing a drawing for CAD, or preparing figures for a patent filing. Requires a PATENTFIG_API_KEY environment variable.
Available today. Use it from your connected AI after setup.
No other account needed.
Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.
Then ask your AI: use the PatentFig AI API skill
What this skill tells your AI
The instructions your AI receives, as published by davila7/claude-code-templates in cli-tool/components/skills/creative-design/patentfig/SKILL.md and read by ahel’s review.
Generate and convert patent figures through the PatentFig AI REST API.
- Base URL:
https://patentfig.ai/api/v1 - Auth:
Authorization: Bearer $PATENTFIG_API_KEY - All endpoints are synchronous and return JSON:
{ "success": true, "data": {...} }on success,{ "success": false, "error": { "code", "message" } }on failure. - Generated files are returned as URLs on
https://cdn.patentfig.ai— pass them to the user or feed them into follow-up calls.
Setup
The key comes from the PATENTFIG_API_KEY environment variable. Never hardcode it, print it, or write it into files or logs.
If the variable is not set, tell the user to create a key at https://patentfig.ai/settings/api-keys (requires a PatentFig AI account; keys start with pfig_) and export it:
export PATENTFIG_API_KEY="pfig_..."
Verify connectivity and credit balance (free call):
curl -s https://patentfig.ai/api/v1/credits \
-H "Authorization: Bearer $PATENTFIG_API_KEY"
# → { "success": true, "data": { "balance": 500 } }
Generate a patent figure (10 credits)
POST /figures — the core endpoint. One prompt → one figure. Views, diagram type, and composition are all controlled through the prompt; for a multi-view set (front/side/top), make one call per view.
curl -s -X POST https://patentfig.ai/api/v1/figures \
-H "Authorization: Bearer $PATENTFIG_API_KEY" \
-H "Content-Type: application/json" \
--max-time 300 \
-d '{
"prompt": "Exploded view of a wireless earbud charging case showing lid, hinge, charging coil, and battery",
"output": "svg",
"labeled": true
}'
output:"png"(raster figure) or"svg"(stroke-based line art, CAD-friendly). Default"png".labeled: trueadds patent-style numeric reference callouts (100, 102, …).referenceImageUrls: up to 4 public image URLs (sketch/photo/prior figure) used as the structural baseline.- Response
data.urlis the figure; withoutput: "svg"the full SVG source is also indata.svg.
Prompting tips: name the parts to draw (names tell the model WHAT to draw, not what to label); for flowcharts/block diagrams, name every node explicitly.
Convert an existing image
All three endpoints accept either JSON with imageUrl (public URL) or multipart/form-data with a file field (PNG/JPEG/WebP/TIFF, ≤10 MB).
Vectorize (20 credits) — raster → SVG / DXF / vector PDF:
curl -s -X POST https://patentfig.ai/api/v1/vectorize \
-H "Authorization: Bearer $PATENTFIG_API_KEY" \
-H "Content-Type: application/json" \
--max-time 300 \
-d '{ "imageUrl": "https://example.com/drawing.png", "format": "dxf", "engine": "lineart" }'
engine: "lineart" (default) redraws as clean single-stroke patent line art — use for CAD/DXF. engine: "trace" reproduces the input pixels faithfully, preserving fills — use when output must match the original.
Enhance (20 credits) — AI super-resolution: POST /enhance with scale (2 or 4, controls pixels) and dpi (300 or 600, metadata stamp only).
Convert (20 credits) — filing-ready raster: POST /convert with format (png/tiff/pdf) and dpi (300 or 600).
Full parameter tables for every endpoint: see references/endpoints.md. Machine-readable spec: https://patentfig.ai/api/openapi.yaml
Operational rules
- Set a client timeout of at least 300 seconds — generation and vectorization take from ~20s up to a few minutes.
- Credits are charged only on success; retrying a failed call never double-bills.
- Rate limit: 60 requests/minute per key. On
429back off (start at 5s) and retry. - On
402 INSUFFICIENT_CREDITS, stop and tell the user to top up at https://patentfig.ai/pricing. - On
500 GENERATION_FAILED, one retry is reasonable; persistent failures usually mean the input doesn't suit the pipeline (e.g., a photo sent for line-art vectorization — suggestPOST /figureswith the photo asreferenceImageUrlsinstead). - Before a batch of billable calls, check
GET /creditsand confirm the balance covers it.
Signals
- GitHub stars
- 32k
- Forks
- 4k
- Last commit
- Sep 2026
Advanced
- Item type
- skill
- Key
patentfig- Source
- github.com/davila7/claude-code-templates