When to use Repurai
Use Repurai when a user wants to search the sources they already follow, review what those sources published today, retrieve saved summaries or YouTube transcripts, or collect Post Automation drafts. It is a user's feed library rather than a general web search API. Cite original post links and ask for permission before refreshing feeds, generating paid AI output, changing saved items or deleting resources.
REST API and authentication
Start with GET https://repurai.com/api/v1 for public discovery. Download /openapi.json for OpenAPI 3.1 operation IDs, parameters and response schemas. Account endpoints require Authorization: Bearer followed by an API key or OAuth access token; API keys are available at /dashboard/api-keys on Starter plans and above. Follow /auth.md and /docs/authentication for credential setup. Never send passwords or credentials in URLs. Plan and monthly usage limits still apply.
Read and search
With the user's credential, GET /api/v1/feeds lists followed feeds. GET /api/v1/posts/today retrieves today's posts and GET /api/v1/posts/search?q=agents searches the library. Paginated lists use page and pageSize (at most 100); continue while hasMore is true. Prefer saved summaries and transcripts before creating new ones. Errors include error, message, status and a resolution hint. A 401 needs a valid credential; a 403 means access or plan limits need attention.
MCP, CLI and automation
MCP clients connect to https://repurai.com/api/mcp. OAuth discovery and sign-in are described in /auth.md. The MCP server card is at /.well-known/mcp/server-card.json. The CLI source is maintained in packages/cli in the Repurai repository; registry publication is pending. Post Automation delivery and webhook configuration are documented at /docs/autopilot. Keep webhook destinations and credentials private, and review publishing permissions before enabling automation.