Introduction
What Pullwire is, how it works, and where to start.
Pullwire gives your AI agent or app live data from the web. You get one API key and one credit balance, and every data tool works the same way: send an input, get clean JSON back, pay only for the results you receive.
How it works
- Pick a tool. Each tool reads one kind of public web data, for example YouTube transcripts, job listings or ads from the Facebook Ad Library. The full list is in the tool reference.
- Start a run. Send the tool's input as JSON to
POST /api/v1/tools/{toolId}/runswith your API key. - Get results. Short runs answer right away with the results. Longer runs can run in the background; you fetch the results when they are done, as JSON or CSV.
- Pay per result. Each result costs the credits listed for its tool. Before a run starts, credits are set aside for the most results it can return; afterwards you are charged only for the results you actually got and the rest goes back to your balance.
Where to go next
Quickstart
Your first API call in a few minutes.
Runs
Sync and async runs, statuses, results as JSON or CSV.
Credits
How credits are reserved, charged and released.
Tool reference
Every tool with its input fields and price.
Connect
MCP: Claude and Cursor
Your agent finds and uses every tool by itself.
TypeScript and Python
Typed SDKs generated from the OpenAPI document.
n8n
One HTTP Request node, imported from cURL.
Make
The HTTP module plus an Iterator.
Zapier
Webhooks by Zapier, Custom Request.
For AI agents
These docs are also available as plain text, so an agent can read them without rendering HTML:
/llms.txt: a short index of the docs and all tools./llms-full.txt: every docs page and the full tool reference in one file./api/v1/openapi.json: the OpenAPI 3 description of the API./api/v1/tools: the live tool catalog as JSON, including each tool's input schema.