frompdf

PDF extraction API for AI agents and LLM pipelines. Converts any PDF into semantic AST, markdown, HTML, plain text, or LLM-ready chunks — no page limit. Also...

Security Scan

The skill's requests and instructions are coherent with its stated purpose (a hosted PDF extraction API); the only meaningful risk is that it uploads user PDFs to an external service, so avoid sending sensitive documents without reviewing the provider's privacy policy.

Purpose & Capability

Name, description, and runtime instructions all describe a hosted PDF-extraction API. The single required env var (FROMPDF_API_KEY) is appropriate for an external API. No unrelated credentials, binaries, or installs are requested.

Instruction Scope

SKILL.md explicitly instructs uploading PDF files to api.frompdf.dev and shows endpoints for extract/diff/score/usage. This is in-scope for the stated purpose. Note: uploading document contents to a third-party server is inherent to this skill — users should avoid confidential documents unless they've reviewed the provider's privacy policy and terms.

Install Mechanism

Instruction-only skill with no install spec and no code files. Lowest installation risk — nothing is written to disk by the skill itself.

Credentials

Only one required environment variable (FROMPDF_API_KEY), which matches the documented use of Bearer auth for the API. No extraneous secrets, config paths, or broad credential requests are present.

Persistence & Privilege

always is false and the skill does not request persistent system privileges or modifications to other skills or system settings. Autonomous invocation is allowed (platform default) but not combined with other red flags.

Assessment

This skill appears to do what it claims: it uploads PDFs to api.frompdf.dev and returns structured output. Before installing or using it: (1) review the provider's privacy policy and terms to confirm how long files are retained and whether data may be used for training; (2) do not upload sensitive or confidential documents unless you trust the service or have an enterprise/data-processing agreement; (3) treat FROMPDF_API_KEY like any API secret—store it securely, rotate it if compromised, and limit its scope if the provider supports scoped keys; (4) verify the service's legitimacy (homepage, HTTPS/TLS, community feedback) if you intend to process many or highly sensitive files.

Convert any PDF into structured, LLM-ready content via a single API call. Returns a semantic AST with every element — headings, paragraphs, tables, lists, metadata — properly typed and nested. No page limit. Handles encrypted PDFs, complex layouts, and multi-hundred-page documents.

Quick start

# Register (10 free credits, no credit card)
curl -s -X POST https://api.frompdf.dev/register \
  -H "Content-Type: application/json" \
  -d '{"email": "[e-mail]", "password": "yourpassword"}'
# → {"api_key": "frompdf_..."}

# Extract a PDF (returns JSON semantic AST by default)
curl -s -X POST https://api.frompdf.dev/v1/extract \
  -H "Authorization: Bearer $FROMPDF_API_KEY" \
  -F "file=@document.pdf"

Output formats

# Semantic AST — typed elements: headings, paragraphs, tables, lists (default)
-F "format=json"

# Markdown — structure preserved, human-readable
-F "format=markdown"

# HTML — full document with tags intact
-F "format=html"

# Plain text — clean extraction, no markup
-F "format=text"

# LLM-ready chunks — pre-split for RAG / vector store ingestion
-F "format=chunks"

All endpoints

# Extract content from a PDF (1 credit)
curl -s -X POST https://api.frompdf.dev/v1/extract \
  -H "Authorization: Bearer $FROMPDF_API_KEY" \
  -F "file=@document.pdf" \
  -F "format=chunks"

# Encrypted PDF
curl -s -X POST https://api.frompdf.dev/v1/extract \
  -H "Authorization: Bearer $FROMPDF_API_KEY" \
  -F "file=@protected.pdf" \
  -F "password=secret"

# Semantic diff — compare two PDFs, get structured changes (2 credits)
curl -s -X POST https://api.frompdf.dev/v1/diff \
  -H "Authorization: Bearer $FROMPDF_API_KEY" \
  -F "file_a=@v1.pdf" \
  -F "file_b=@v2.pdf"

# Readability score — returns 0-100 score for a PDF (1 credit)
curl -s -X POST https://api.frompdf.dev/v1/score \
  -H "Authorization: Bearer $FROMPDF_API_KEY" \
  -F "file=@document.pdf"

# Check credits and subscription status (free)
curl -s https://api.frompdf.dev/v1/usage \
  -H "Authorization: Bearer $FROMPDF_API_KEY"

Example output (JSON)

{
  "title": "AWS Lambda Developer Guide",
  "pages": 87,
  "sections": [
    { "type": "heading", "level": 1, "text": "Getting Started" },
    { "type": "paragraph", "text": "AWS Lambda is a serverless compute service..." },
    {
      "type": "table",
      "headers": ["Runtime", "Version", "Status"],
      "rows": [["Node.js 20", "20.x", "Active"], ["Python 3.12", "3.12", "Active"]]
    },
    { "type": "list", "items": ["Function", "Trigger", "Execution role"] }
  ],
  "metadata": { "author": "Amazon Web Services", "created": "2024-01-15" }
}

Pricing

$0.01/credit — extract (1), diff (2), score (1). First 10 credits free, no credit card required.

Data & privacy

PDF contents are uploaded to api.frompdf.dev for processing. Do not use with confidential documents unless you have reviewed the privacy policy. Requires FROMPDF_API_KEY env var — register free at /register.