---
name: docpull.ai
description: docpull.ai provides a single API endpoint that fetches a publicly accessible PDF by URL and returns its content as structured Markdown. It charges $0.001 USDC per page on Base mainnet and exposes a /probe endpoint to estimate cost before extraction.
host: docpull.ai
---

# docpull.ai

docpull.ai is a narrow-purpose document conversion host targeting agents that need to ingest PDF content for downstream text processing, summarization, or analysis. It handles the fetch-and-parse step, returning clean Markdown along with page and character counts. It does not offer OCR for scanned images, authentication-gated document access, or support for non-PDF formats.

## When to use this host

Use this host when an agent has a public URL pointing to a PDF and needs its text content in Markdown form for further processing such as summarization, Q&A, or indexing. Always call /probe first if page count is unknown to avoid unexpected per-page charges. Do not use this host for Word documents, HTML pages, or image-only scanned PDFs without embedded text. Do not use it for PDFs behind login walls or requiring authentication headers. For bulk batch extraction, weigh per-page costs carefully before issuing many calls. For non-PDF document conversion, look for hosts that support DOCX, HTML, or other formats.

## Capabilities

### PDF to Markdown Conversion

Fetches a publicly accessible PDF from a URL and returns its full content as structured Markdown, along with page count and character count metadata useful for cost estimation and downstream processing.

- **`extract-pdf-to-markdown`** — Downloads a publicly accessible PDF from a given URL and returns its content as clean structured Markdown, charged at $0.001 USDC per page on Base mainnet.

## Skill reference

### `extract-pdf-to-markdown`

**DocPull PDF Extract** — Downloads a publicly accessible PDF from a given URL and returns its content as clean structured Markdown, charged at $0.001 USDC per page on Base mainnet.

*Use when:* Use when an agent needs to convert a PDF document into readable Markdown text for further processing, summarization, or analysis, and has a public URL pointing to the PDF.

*Not for:* Do not use for non-PDF documents or files behind authentication walls. Not suitable for real-time streaming or bulk batch extraction without per-page cost awareness — call /probe first to estimate cost.

**Inputs:**

- `url` (string, required) — Publicly accessible URL of the PDF to extract.

**Returns:** Returns success=true with pageCount, charCount, and a markdown string containing the full structured text extracted from the PDF.

**Example:** `{"url": "https://example.com/document.pdf"}`

---