Teach any AI agent how to use the DeepRead OCR API.
Install these skills and your AI coding assistant instantly knows how to integrate DeepRead — sign up, get an API key, extract text and structured data from documents, fill PDF forms, redact PII, use blueprints for optimized accuracy, handle webhooks, and manage errors.
| Skill | What it does |
|---|---|
/deepread-setup |
Get started — signup, API key, first request, structured extraction, blueprints |
/deepread-api |
Full API reference — all endpoints, auth, schemas, blueprints, webhooks, code examples |
/deepread-form-fill |
Fill PDF forms — submit form + JSON data, get back completed PDF with quality report |
/deepread-pii |
PII removal (Enterprise plans) — upload documents, detect and redact personal information, download clean files |
npx skills add deepread-tech/skillsAfter installing, use the slash commands:
/deepread-setup # walks you through signup, API key, and first request
/deepread-api # full API reference — your agent can write integration code in any language
/deepread-form-fill # fill PDF forms with AI — any PDF, no AcroForm needed
/deepread-pii # detect and redact PII from documents
Copy the rule files into your project:
git clone https://github.com/deepread-tech/skills.git /tmp/deepread-skills
cp -r /tmp/deepread-skills/.cursor .Or manually copy .cursor/rules/deepread-setup.mdc, .cursor/rules/deepread-api.mdc, .cursor/rules/deepread-form-fill.mdc, and .cursor/rules/deepread-pii-removal.mdc into your project's .cursor/rules/ directory.
Rules use alwaysApply: false — Cursor will invoke them automatically when you're working with the DeepRead API.
Copy the rule files into your project:
git clone https://github.com/deepread-tech/skills.git /tmp/deepread-skills
cp -r /tmp/deepread-skills/.windsurf .Or manually copy .windsurf/rules/deepread-setup.md, .windsurf/rules/deepread-api.md, .windsurf/rules/deepread-form-fill.md, and .windsurf/rules/deepread-pii-removal.md into your project's .windsurf/rules/ directory.
clawhub install uday390/deepread-ocrThe skills are published to the OpenClaw registry as user-invocable commands. Use /deepread-setup, /deepread-api, /deepread-form-fill, and /deepread-pii after installation.
AI-native document processing API. Upload a PDF or image, get back extracted text and structured data. Two engines: Extract (one OCR pass) and Deep Extract (two OCR passes and an LLM judge). Both flag uncertain fields for human review; accuracy is measured on public benchmarks at deepread.tech/benchmarks.
Key features:
- OCR — Structured extraction via JSON Schema
- Form Fill — AI-powered PDF form filling with visual QA
- PII Removal — Detect and redact 14 types of personal information (Enterprise plans)
- Blueprints — optimized reusable schemas
- Per-field review flags (
needs_review+review_reason) for human-in-the-loop review - Human review in DeepRead — send
reviewer_emailand a person checks and corrects the result - Webhook notifications
- Shareable preview links (no auth needed)
Pricing: Extract $15 and Deep Extract $35 per 1,000 pages; the Free plan includes 2,000 pages a month.
Links:
- Dashboard:
https://www.deepread.tech/dashboard - Docs:
https://www.deepread.tech/docs
MIT — DeepRead Technologies