Connect to your Twin
Bring approved knowledge from your Twin into the tools where you already work.
You do not need to learn a new technical system to start using your Twin. You can use the Twin workspace directly, connect an assistant you already use, or add approved Twin knowledge to your own software.
Let your agent set it up
Copy one prompt and let your agent identify and configure the right connection.
Choose where to work
Compare ChatGPT, Claude, Codex, Claude Code, the Twin workspace, and developer options.
Use the Twin CLI
Sign in, inspect channels, query your Twin, and submit feedback from a terminal.
Check availability
Separate user interfaces, configured model providers, and routes enabled for your organisation.
Understand access
See what is shared, how sign-in works, and what your assistant cannot request.
Fix a connection
Diagnose sign-in, channel access, timeout, and configuration problems.
What a connection gives you
A connected assistant can help with a bounded read journey:
- List the channels available to you.
- Inspect a channel's queryable object and field schema.
- Read its recent activity and, with explicit approval, publish a post or reply.
- Ask questions through the same router as the Twin workspace.
- Submit typed deterministic channel queries when an exact result shape is needed.
- Submit feedback that you provide or confirm, optionally attached to a routed query.
The assistant cannot choose a different tenant, inspect source rows, write SQL, change the model, approve business meaning for you, or bypass channel access. The service derives your organisation from sign-in and keeps data bindings and routing policy under server control.
Availability
Connections are enabled per organisation. Ask your implementation contact for a connection pack before you begin. It confirms that the organisation is enabled and records any setup route specific to your tool.
Start with a copied prompt
The quickest route is the agent-assisted setup prompt. Choose the short hosted-assistant or coding-agent version. It tells the agent to:
- follow the right tool-specific guide;
- preserve existing configuration;
- keep credentials out of chat, files, commands, and logs;
- stop when an organisation owner must enable access;
- verify the connection with a natural question; and
- report exactly what changed.
If you prefer to work through the setup yourself, choose your tool.