Giving AI Agents a Body.
Self-developed Skills and REST APIs for platforms and enterprises in the agent era.
中文 · Skills · REST API · Agent Quickstart
All capabilities are built, tested, and operated by Agent Body, so platforms do not have to evaluate and maintain inconsistent community services.
- Built in-house: We develop every Skill and its backend service.
- Reliable: Thoroughly tested and continuously maintained.
- One API surface: Use fixed REST APIs with one account key.
- Dedicated support: One-on-one integration support for enterprise customers.
Skills teach an AI agent the workflow, input limits, safety boundaries, and result handling for a capability. The agent calls Agent Body's fixed REST APIs directly.
Tell your AI agent:
Install all official Agent Body Skills from https://github.com/agentbody/skills. Keep the Skill files unchanged. Read https://agentbody.io/docs/agent-quickstart.md and use only Agent Body REST APIs.
For one workflow, use its source directory:
Install the official Agent Body document-parsing Skill from https://github.com/agentbody/skills/tree/main/skills/document-parsing. Keep the Skill file unchanged and use the current Agent Body REST API contract.
| Skill source | Current REST API | Example request after installation |
|---|---|---|
| account-usage | GET /v1/account/quota, GET /v1/usage/summary, GET /v1/usage/history |
"Show my current quota and recent usage." |
| people-data | GET /v1/linkedin/profile, /email, /phone, GET /v1/youtube/channel/email |
"Find public business email data for this profile." |
| document-parsing | POST /v1/documents/parse, GET /v1/documents/{document_id} |
"Parse this HTTPS PDF and return pages 1 through 5 as Markdown." |
| humanize-writing | POST /v1/text/humanize |
"Rewrite this text naturally while preserving all facts." |
| youtube-transcript | GET /v1/youtube/transcript |
"Extract the English captions from this YouTube video." |
| tiktok-transcript | GET /v1/tiktok/transcript, POST /v1/tiktok/transcribe |
"Extract existing captions from this TikTok video." |
Production base URL: https://api.agentbody.io
Every API request uses a Bearer key and snake_case field names. Store the key in the agent runtime's secret store or AGENTBODY_API_KEY; never commit, log, or expose it to browser code.
curl -sS https://api.agentbody.io/v1/account/quota \
-H "Authorization: Bearer ${AGENTBODY_API_KEY}"Use the quota endpoint as the first read-only verification. For the complete API catalog, request examples, API-key safety rules, and a copyable agent prompt, read:
The public API contract contains the following fixed routes:
GET /v1/account/quota
GET /v1/usage/summary
GET /v1/usage/history
GET /v1/linkedin/profile
GET /v1/linkedin/email
GET /v1/linkedin/phone
GET /v1/youtube/channel/email
GET /v1/youtube/transcript
GET /v1/tiktok/transcript
GET /v1/documents/{document_id}
GET /v1/seo/google-ads-search-volume
POST /v1/documents/parse
POST /v1/text/humanize
POST /v1/tiktok/transcribe
For billable POST requests, use a unique Idempotency-Key per logical request. Account, usage, and document-result reads are free. Successful responses return business JSON directly; errors use {"error":{"code":"...","message":"..."}}.