Skip to main content

API Overview

The API exposes three ways to convert a document to Markdown, plus a state endpoint. Everything answers text/markdown by default, or JSON if you order it.

Base URL

Authentication

All calls use x-api-key. See Authentication.

Endpoints

POST /convert

Upload the file as multipart/form-data.

POST /convert/raw

Send the raw binary in the body. Supports gzip and zstd.

POST /convert/from-url

Convert from a signed URL from S3, R2, Supabase, GCS, or Azure.

GET /jobs/{id}

Status of an automatically queued conversion (202).

GET /

Service status and supported formats.

Parameters

Complete reference of query params, modes and formats.

Answer

Observability headers and body JSON.

The three conversion routes

For large documents, from-url usually wins: your server doesn’t receive or forward the same file twice.

Response model

By default you receive raw Markdown (text/markdown; charset=utf-8) plus observability headers. With response_format=json you receive an object with Markdown and metadata embedded. See Response and headers.

Status codes

Detail in Errors.