!

Legal Disclaimer

PipeAgent is a data distribution gateway. We do not own, verify, or endorse the data provided by third-party creators. Use at your own discretion.

Docs/consumer / quickstart

Quickstart

From API key to your first JSON response—follow these steps in order.

1. Get an API key

See Authentication for how to create a read key and how to send it via x-api-key or Authorization: Bearer. Store keys in environment variables or a secret manager—never commit them.

2. Pick a feed

Open the Feeds directory in the app, choose a feed, and copy its Feed ID from the detail page.

To discover feeds by keyword before you commit to an ID, use the search endpoint (same pattern as *Discovery* in Agent Integration). Discovery requires your read key (limit / offset paginate results; default limit is 20, max 100):

bash
curl -s "https://api.pipeagent.dev/v1/search?q=YOUR_KEYWORDS&limit=20&offset=0" \
  -H "x-api-key: YOUR_API_KEY"

For a full directory listing, use GET https://api.pipeagent.dev/v1/catalog with the same headers and pagination. To inspect type and schema without pulling feed rows, use GET https://api.pipeagent.dev/v1/feed/{FEED_ID}/metadata.

Consumer read APIs (feed fetch, search, and related GET routes) live on https://api.pipeagent.dev with paths /v1/... — for example /v1/feed/{id} and /v1/search. The site https://pipeagent.dev is the web app and dashboard, not this JSON API host.

3. Make your first fetch

Consumers read data with GET (singular feed):

GET https://api.pipeagent.dev/v1/feed/{FEED_ID}

bash
curl -s "https://api.pipeagent.dev/v1/feed/FEED_ID" \
  -H "x-api-key: YOUR_API_KEY"

Bearer style (same as Authentication):

bash
curl -s "https://api.pipeagent.dev/v1/feed/FEED_ID" \
  -H "Authorization: Bearer YOUR_API_KEY"

A successful response is JSON. For 401 / 402 / 403 and other codes, see the error table in Usage Limits.

4. Optional query parameters

To shrink payloads and save LLM tokens, add (details in Usage Limits):

  • jsonpath — filter JSON at the edge before it reaches your agent.
  • limit / offset — pagination; limit is capped at 100 per request.
  • cursor — cursor-based pagination for collection and stream feeds.
  • start_time / end_time — for stream feeds only; inclusive time range on event_timestamp (ISO-8601).
  • Example:

    bash
    curl -s "https://api.pipeagent.dev/v1/feed/FEED_ID?limit=10&offset=0&jsonpath=%24%5B%2A%5D.title" \
      -H "x-api-key: YOUR_API_KEY"

    Stream time-window example:

    bash
    curl -s "https://api.pipeagent.dev/v1/feed/FEED_ID?limit=20&start_time=2026-03-24T00:00:00Z&end_time=2026-03-24T23:59:59Z" \
      -H "x-api-key: YOUR_API_KEY"

    5. OpenClaw, LangChain, CrewAI

  • OpenClaw — Overview: OpenClaw Integration. Conversation flow: PipeAgent Pilot (official SKILL.md). Terminal install: CLI install. On any feed page, the OpenClaw tab copies the one-line install.
  • LangChain / CrewAI — See the framework guides linked from Agent Integration.
  • 6. Next steps

  • API for AI Agents — single-page reference for agents (read, limits, errors)
  • Pricing & Billing — free feeds vs metered paid usage
  • Feed types — how Singleton, Collection, and Stream behave when you read data
  • Machine-readable overview: /openapi.json
  • Version 1.0.4 - Premium Infrastructure