Announcing the Growvo Public API and hosted MCP server

From day one, Growvo has been the place where content gets made and published. Today it also becomes a platform you can program. The Growvo Public API is live for every workspace, and alongside it we are shipping growvo-mcp — a Model Context Protocol server that exposes the API as tools for Claude, Cursor, or any MCP-capable agent.

A real API, not an export button

Create a key in Settings → API keys and you can drive the same pipeline the app uses: publish to any connected network, schedule calendar events, register media from public URLs, read your saved outputs, and pull post analytics with engagement scores. Every request acts as the user who created the key, and the API re-checks workspace membership and role permissions live on every call — revoke a key or demote a user and the change takes effect immediately.

curl -H "Authorization: Bearer REDACTED" \
  https://api.growvo.ai/v1/me

What you can build with it

• Publish from your own product: POST /v1/posts crossposts to up to 13 networks in one call, with per-network text variants and native thread support.
• Automate your calendar: GET /v1/calendar/next-slot finds the next free slot, POST /v1/schedule books it.
• Feed your prompts: POST /v1/sources/ingest fetches any public URL and returns clean plaintext, SSRF-guarded.
• Generate media: POST /v1/images and POST /v1/videos create images and faceless-video storyboards straight into your library.

Your AI agent gets hands

The MCP server is a thin, stateless bridge: point it at your API key and every tool call becomes an authenticated request to the live API. Ask Claude to “list my connected accounts and publish this thread to X and Mastodon” and it happens — with the same permissions, quotas, and audit surface as the app.

{
  "mcpServers": {
    "growvo": {
      "command": "npx",
      "args": ["-y", "@growvo/mcp-server"],
      "env": { "GROWVO_API_KEY": "REDACTED" }
    }
  }
}

Rate limits are boring on purpose: 60 requests a minute per key, honest 429s with Retry-After, and plan quotas that behave exactly like the in-app ones. The full reference lives in our GitHub docs — start at /developers for the quickstart.