Yes. OneUp works from the terminal through coding agents like Claude Code, Codex, and Cursor, connected over MCP. You add OneUp with one command, approve access in your browser, and then schedule and manage posts by describing what you want.
Connect OneUp from your terminal
Claude Code
claude mcp add --transport http --scope user oneup "https://feed.oneupapp.io/mcp/oneup"
Then open Claude Code, run /mcp, select oneup, and authenticate.
Codex
codex mcp add oneup --url "https://feed.oneupapp.io/mcp/oneup"
If authentication doesn't start automatically, run codex mcp login oneup. Use codex mcp list to confirm the setup.
Cursor
Add OneUp to ~/.cursor/mcp.json:
{
"mcpServers": {
"oneup": {
"url": "https://feed.oneupapp.io/mcp/oneup"
}
}
}
How sign-in works
Generate an API key on OneUp's API Access page. When you authenticate, OneUp opens a browser page where you enter the key and approve the agent's access. OneUp stores the key encrypted, and the agent receives an OAuth token instead of the key itself. You can review and revoke connections from OneUp's MCP Connections page.
Some clients can only connect with a key in the URL. If you use that option, keep the full URL private, because it contains your credential and may be saved in your client's configuration or shell history.
Check the connection
Start a session and ask: "Use OneUp to list my connected social accounts." If the agent lists them, you're set.
What you can do
- Create text, image, and video posts, and upload media
- List, edit, delete, and reschedule scheduled, draft, published, and failed posts
- Fetch and reply to comments
- Read and send inbox messages
- Pull analytics across supported platforms
Is it a traditional CLI?
Not in the sense of fixed subcommands and flags. You describe what you want and the agent runs it. If you'd rather write your own scripts, OneUp also has a documented REST API with API key authentication, and the endpoints and required fields are in the reference docs.
Ready to try it? See OneUp's coding agent setup guide.