Skip to main content
CaptureFlow turns your expertise (videos, voice notes, documents, links) into publish-ready posts. The API and the MCP server give AI assistants and your own tools the same building blocks you use in the app:
  • Read your ideas, inputs, posts, content strategy and connected social accounts.
  • Create ideas, inputs (text, links, files) and draft posts.
  • Propose a post for scheduling or posting. You accept or reject it in CaptureFlow.
  • Start the content agent, which writes drafts from an idea, an input or text.
The API never publishes. Scheduling or posting is always a proposal that you accept in CaptureFlow. An AI assistant can prepare everything; you decide what goes out.

​
Two ways in

MCP server

For AI assistants: Claude, ChatGPT, Claude Code, Cursor, VS Code and more. Add https://mcp.captureflow.ai/mcp, sign in, and choose what the assistant may do.

REST API

For scripts, automations and your own apps. https://api.captureflow.ai/v1, with an API key or OAuth 2.1.
Both expose the same operations, with the same permissions, limits and errors. Each MCP tool is named after its API operation (for example create_post_draft is POST /v1/posts).

​
Who can use it

The API and MCP server work in any workspace with a paid plan or a trial. An owner or admin can switch API access off for a workspace in Settings > API & MCP.

​
Next steps

Quickstart

Create a key and make your first calls.

Authentication

API keys and connected apps.

Approvals

How proposals reach you.

Errors

Every error code, with what to do.