LLM Providers

LLM Providers

LLM providers power prompt-based text processing, transcription cleanup, and memory extraction. TypeWhisper for Linux bundles several providers, and each one has its own per-provider Stream responses toggle (on by default) that streams LLM output token by token into the dictation overlay and prompt palette.

Bundled providers

Provider	Notes
Claude	Cloud provider.
OpenAi	Cloud provider.
OpenAiCompatible	Route for a local or LAN OpenAI-compatible server (for example Ollama).
OpenRouter	Cloud provider.
Gemini	Cloud provider.
GemmaLocal	Local option.
Groq	Cloud provider.
Cerebras	Cloud provider.
Cohere	Cloud provider.
Fireworks	Cloud provider.
Xai	Cloud provider.

Each provider exposes a per-provider Stream responses toggle, on by default, controlling token-by-token streaming of responses into the dictation overlay and the prompt palette.

Local and LAN options

• OpenAiCompatible is the route for a local or LAN OpenAI-compatible server, such as Ollama. This lets the cleanup and prompt pipeline run against a model you host yourself.
• GemmaLocal is a local provider option.

Related pages

• Prompts — prompt actions, provider overrides, and the prompt palette that LLM providers drive.
• Text cleanup — the Medium/High cleanup levels route through the configured LLM provider.
• Memory — long-term memory extraction depends on a configured LLM provider.

---

Changelog

Date	Change
2026-06-17	Initial version.