One workflow. Every surface.
Schedule from your terminal, from Claude or Cursor over MCP, from the typed API, or from the dashboard. Same auth, same drafts, same queue, wherever you are.
- CLI and MCP server, first-class
- Typed API with a generated SDK
- Structured output agents can chain
What you can do.
A CLI agents can read
Run octospark post from your terminal. Output is structured, so agents can chain it without scraping stdout.
MCP server for Claude and Cursor
Point any MCP client at OctoSpark. List drafts, queue posts, and read schedules without leaving the editor.
Typed API and SDK
Every action is a typed endpoint with a generated SDK. Build your own surface on top with full contract safety.
The dashboard, for humans
Prefer a calendar and a click? The UI runs on the exact same API and queue as the CLI and your agents.
How it works.
- 1
Authenticate once
Connect your account and mint a key. The same identity works on every surface.
- 2
Pick your surface
Terminal, agent, API, or dashboard. They all read and write the same drafts and queue.
- 3
Ship
Queue a post anywhere and it shows up everywhere, instantly.
Frequently asked questions.
Yes. The CLI emits structured output designed for agents to parse, so Claude or Cursor can chain commands without guessing.
Any MCP-compatible client, including Claude Code and Cursor. Run the server locally or remote and point your client at it.
Yes. CLI, MCP, API, and dashboard read and write the same drafts, schedule, and analytics. There is one source of truth.
Yes. The API ships a generated, type-safe SDK so you can build your own surface with full contract safety.
Schedule from wherever you work.
Get early access to the agent-native scheduler built for developers who ship.
Get early access