Providers
Configure a provider
- Open Settings in the Nakama dashboard
- Go to LLM providers
- Add a provider and paste your API key
- Browse or enter models for that provider
- Assign a model to each profile that should use it
On first run, Nakama can also prompt for a provider during the setup wizard. See First-time setup.
Settings are saved in ~/.nakama/config.ini (or under NAKAMA_CONFIG_DIR when set).
During First-time setup, the wizard uses the same provider form:
Supported providers
Nakama supports these built-in provider types:
| Provider | Notes |
|---|---|
| OpenAI | GPT models; required for OpenAI audio transcription (Telegram voice notes) |
| Anthropic | Claude models |
| OpenRouter | Route to many models through one API key |
| Gemini | Google Gemini models |
| DeepSeek | DeepSeek models |
| Cerebras | Cerebras models |
| Cloudflare Worker AI | Workers AI models; add an API key and account ID in Settings (saved to config.ini). Llama 3.1 8B is the cheaper catalog option. CLOUDFLARE_ACCOUNT_ID is a fallback only |
| Fireworks | Fireworks AI models |
| Ollama | Local or Ollama Cloud; multiple instances allowed |
| OpenCode Go | OpenCode Go endpoint |
| Custom (OpenAI-compatible) | Any OpenAI-compatible API; multiple instances allowed |
Most built-in types allow one configured instance each. Ollama and Custom (OpenAI-compatible) can be added multiple times (for example local Ollama plus a remote compatible endpoint).
Provider-specific features
Some capabilities depend on the active provider:
- Web search (
web_searchtool) — OpenAI or Anthropic with a valid API key; not available on OpenRouter - Audio transcription (Telegram voice notes) — requires an OpenAI provider and a transcription model in Settings → Audio transcription model
- Image parsing — used when the chat model cannot see images. Set Settings → Image parsing model. The list only includes models marked as vision-capable. For a custom endpoint, turn Vision on for that model under Settings → LLM providers
- Coding agent provider passthrough — spawn-time credentials for Claude Code / Codex / OpenCode; see Coding agent
Per-profile models
Each profile selects its own model from the configured providers. Two profiles in the same org can use different providers or models.
Platform admins manage provider credentials globally. Org members use the models assigned to their profiles.
Next steps
- First-time setup — complete the setup wizard
- Profiles — assign models and tools per bot
- Builtin tools — provider-dependent tool availability