-
Notifications
You must be signed in to change notification settings - Fork 0
Setting up an assistant provider
Lwando Zwane edited this page Aug 18, 2026
·
1 revision
The assistant talks to a model provider you pick in Preferences → Assistant.
| Provider | Key to set | Notes |
|---|---|---|
| Kimi (Moonshot) | MOONSHOT_API_KEY |
default |
| GLM (Zhipu) | ZHIPU_API_KEY |
|
| Claude (Anthropic) | ANTHROPIC_API_KEY |
|
| DeepSeek | DEEPSEEK_API_KEY |
Pro or Flash |
| Requesty (router) | REQUESTY_API_KEY |
routes Claude Opus |
| Custom | — | any OpenAI/Anthropic-compatible endpoint |
Tome reads keys from your login shell (export them in your shell profile and restart Tome). The key stays in the main process and never reaches the browser layer. A custom provider's key is stored locally in the app's 0600 store.
- Custom provider (Preferences → Assistant) — set base URL, model, key, and wire (OpenAI or Anthropic).
-
Env override —
TOME_CHAT_BASE_URL/TOME_CHAT_MODELpoint the assistant at any OpenAI- or Anthropic-compatible endpoint (including a local one).
The assistant's error names exactly which key to set. Export it in your shell, restart Tome, and try again.
Tome on GitHub · Report issues in the repository, not the wiki.