BiGapiguides
POST /v1/pdf/pages → 200 image/jpeg · 1¢

PDF to images via API

Rasterize pages to JPEG (default) or PNG, 36–600 dpi. One page → one image, several pages → a ZIP. Built for vision models that want to look at documents.

Quick start (curl)

Get a key first – no signup: curl -X POST https://api.bigapi.dev/v1/keys (100 free operations, 7 days).

curl -X POST https://api.bigapi.dev/v1/pdf/pages -H 'Authorization: Bearer $KEY' -F 'file=@in.pdf' -F 'first=1' -F 'last=1' -F 'format=png' -o page1.png

Give it to your agent (MCP)

Add BiGapi to Claude Desktop, Cursor or any MCP client – the server fetches its own key on first use:

{
  "mcpServers": {
    "bigapi": { "command": "npx", "args": ["-y", "@bigapi/mcp"] }
  }
}

Then simply ask: “Convert page 1 of this PDF to a JPEG so you can look at it.” – the model calls the pdf_to_images tool.

When your agent needs this

Parameters

ParamTypeRequiredNotes
filefileyes
dpiintegerno36–600 – default 150
firstintegernoFirst page (1-based).
lastintegernoLast page.
formatstringnoone of jpeg, png · default jpeg
qualityintegernodefault 85

Pricing

1¢ per operation
First 100 operations free (per key, 7 days) · failed calls (4xx/5xx) cost $0 · repeats with Idempotency-Key are free · prepaid credit from $5, never expires.

Every response carries X-BigAPI-Cost, X-BigAPI-Balance and X-BigAPI-Free-Ops – your agent can budget on its own.

Errors your agent can handle

StatusMeaningAgent action
401missing/invalid keyPOST /v1/keys for a fresh one
402credit exhaustedshow the upgrade_url from the response to a human
413file too largesplit the input (limit 100 MB)
422conversion failedcheck the detail field – costs $0
429rate limitedwait retry_after seconds

FAQ

JPEG or PNG for vision models?

JPEG at 150 dpi is the sweet spot – 10× faster to encode and plenty sharp for reading.

How do first/last work?

first=2&last=4 renders pages 2–4 only.

What do I get for multi-page input?

A ZIP with page-001.jpg, page-002.jpg, …