@devries-cameron20 / t3code Mirror · read-only
Fork 0

The best way to control Claude Code, Codex and more from any computer or phone

Mirrored from https://github.com/Camo67/t3code.git

Synced Nov 21, 03:37 UTC

main
md 42 lines 2.31 KB
Rendered Raw
1 # Text generation models
2
3 T3 Code writes some text for you: thread titles, commit messages, pull request titles and
4 descriptions, branch names, and voice transcript cleanup. Pick the model for this in
5 **Settings → Threads → Text generation model**. **Settings → Source control & issues** and
6 **Settings → Voice** can each pick a different one.
7
8 By default, a coding agent such as Codex or Claude writes this text, starting a session for each
9 request. A **text generator** — a model API that only writes text — is usually faster and cheaper.
10
11 ## Add a text generator
12
13 Go to **Settings → Providers → Text generators** and choose **Add generator**:
14
15 - **OpenAI** or **Anthropic** — enter an API key.
16 - **OpenRouter** — enter an OpenRouter key to use any model it serves.
17 - **Custom (OpenAI-compatible)** — enter the base URL of any other endpoint, such as Ollama at
18 `http://localhost:11434/v1` or LM Studio. Local endpoints usually need no key. The environment's
19 server makes the requests, so on a remote environment `localhost` means that machine, not yours.
20
21 Give it a name if you add more than one from the same service. Your key is stored on the
22 environment, like other [secrets](./secrets.md), and can come from a secret provider instead of
23 being typed in.
24
25 T3 Code lists the service's models to check the connection, then shows the model picker for each
26 text task so you can switch any of them to the new generator, or keep what they use. **Test** next
27 to a picker has the chosen model answer a tiny prompt, which proves the key works; some services
28 list their models without checking it. **Show models** in the generator's menu lists its models
29 and refreshes the list after the service adds some.
30
31 Each request sends the API what the task needs, such as the staged diff for a commit message.
32
33 Turning a generator off or removing it sends its work back to the environment's text generation
34 model.
35
36 ## What text generators can't do
37
38 Review guides read the checkout over many steps, so they always use a coding agent. If your source
39 control writer model is a text generator, review guides use Codex or Claude instead, and fail if
40 neither is set up. To choose which, pick one in **Settings → Source control & issues → Review guide model**.
41
42 A [managed policy](./managed-policy.md) that limits providers also turns off text generators.