Use Tether Studio from Claude, ChatGPT, Grok, Gemini or Codex

Tether Studio has its own MCP server, so the AI assistant you already use for research and writing can work in your studio: draft posts in Content Studio, save links and headlines into Resources, read your brand kit, list your phones and start flows. It can't approve, post, send messages or tap on a phone: every run it starts waits for you to approve it in Tether Studio first.

The address

Every app below needs the same server URL: https://tether-studio.app/api/mcp (also shown in Settings → Connected apps). It uses MCP's Streamable HTTP transport.

There are two ways to sign in:

• Sign in with Tether Studio (OAuth). The app opens a Tether Studio page where you pick the studio and tick what it may do. This is the best choice for Claude, ChatGPT and most apps.
• A personal access token. For apps and scripts that take a token instead. Make one in Settings → Connected apps → New access token, tick what it may do, and send it as the header Authorization: Bearer <token>. It is shown once and expires after 30, 90 or 365 days.

Either way the app acts as you, in one studio, with at most what your role there allows (a Viewer's apps can only read). You can disconnect any of them in Settings → Connected apps; owners and admins see the whole studio's.

Claude (claude.ai and the Claude apps)

1. Open Customize → Connectors, choose Add, then Add custom connector. On a Team or Enterprise plan, an owner adds it first under Organization settings → Connectors.
2. Name it Tether Studio and paste the server URL. Leave the client ID and secret empty.
3. Choose Connect, sign in to Tether Studio, pick your studio and choose Allow.

Claude Code: run claude mcp add --transport http tether-studio https://tether-studio.app/api/mcp, then type /mcp in Claude Code and sign in.

ChatGPT

1. On chatgpt.com (the web app), open Plugins, choose +, then Add custom MCP server.
2. Paste the server URL and choose OAuth for authentication.
3. Choose Create as a plugin, install it, sign in to Tether Studio and choose Allow.
4. In a chat, mention it with @ to use it.

Which ChatGPT plans and workspaces can add custom MCP servers is set by OpenAI and your workspace admin. ChatGPT asks you to confirm before it uses a tool that changes something.

Grok

1. On grok.com, open Connectors, choose New Connector, then Custom.
2. Paste the server URL and choose how it signs in. If it offers OAuth, sign in to Tether Studio; if it asks for a token or header, use a personal access token. On Grok Business or Enterprise, an admin adds the connector first in the xAI console.

Grok CLI: grok mcp add --transport http tether-studio https://tether-studio.app/api/mcp.

Codex (OpenAI's coding agent)

• Sign in: codex mcp add tether-studio --url https://tether-studio.app/api/mcp, then codex mcp login tether-studio.
• Or with a token: in ~/.codex/config.toml add [mcp_servers.tether-studio] with url = "https://tether-studio.app/api/mcp" and bearer_token_env_var = "TETHER_STUDIO_TOKEN", and put your access token in that environment variable.

Gemini

• Gemini CLI: gemini mcp add --transport http tether-studio https://tether-studio.app/api/mcp, then /mcp auth tether-studio in a session to sign in. With a token instead, add to ~/.gemini/settings.json under mcpServers: "tether-studio": { "httpUrl": "https://tether-studio.app/api/mcp", "headers": { "Authorization": "Bearer <token>" } }.
• The Gemini app (gemini.google.com): we couldn't find a way to add your own MCP server there yet. Use the Gemini CLI, or prepare your research in Gemini and save it with any of the other apps.

Cursor and VS Code

• Cursor: in mcp.json, add "tether-studio": { "url": "https://tether-studio.app/api/mcp" } under mcpServers. Cursor offers to sign in; or add "headers": { "Authorization": "Bearer ${env:TETHER_STUDIO_TOKEN}" }.
• VS Code: in .vscode/mcp.json, add "tether-studio": { "type": "http", "url": "https://tether-studio.app/api/mcp" } under servers. VS Code offers to sign in.

Any other MCP app

Anything that speaks MCP over Streamable HTTP works: give it the server URL. Apps that sign in find everything themselves (OAuth 2.1 with PKCE, discovery at https://tether-studio.app/.well-known/oauth-protected-resource/api/mcp, dynamic client registration or a client ID metadata document). Apps that can't sign in can send a personal access token.

What you can ask

• "Read my brand kit and draft 3 TikTok posts from these headlines." The drafts land on the Content board as Draft; you edit and approve them as usual.
• "Save these 10 headlines to my daily resources." A scheduled flow with the Saved resources step takes the new ones each morning, has AI build posts from them and sends them on.
• "Run my Reels flow on the Galaxy A07 with this caption." The run waits for your approval under Runs before anything happens on the phone.

Start a flow from an assistant, Zapier or n8n

Add the trigger An AI assistant sends something to a flow and list the fields it takes (e.g. headlines, tone). Assistants see it with its fields and can send to it. For tools that can't speak MCP, turn on Also accept a signed link: the flow gets its own URL and secret, and each request is signed with an HMAC of the timestamp and body, works once and expires after 5 minutes. Either way, the run waits for your approval before any phone step, and every post or message asks again.

Safety

• Every change an app makes is in the Audit log with the app's name.
• Text from your resources and content is treated as data: assistants are told not to follow instructions inside it.
• Apps are limited in how often they can call (per minute and, for downloads, per hour).
• Downloads by link fetch public https addresses only, with the same file types and sizes as uploads.