provider/

Package anthropic is a streaming LLM service backed by the Anthropic API.

Package assemblyai is a streaming speech-to-text service backed by AssemblyAI's Universal-Streaming (v3) WebSocket.

Package azureopenai provides an LLM service for Azure OpenAI.

Package azurespeech provides Azure AI Speech services.

Package bedrock provides a streaming LLM service for Anthropic Claude models served through Amazon Bedrock.

Package cartesia is a streaming text-to-speech service backed by Cartesia's TTS WebSocket.

Package cerebras provides Cerebras's OpenAI-compatible LLM service.

Package deepgram provides Deepgram's streaming speech-to-text service (over the live transcription WebSocket) and its Aura text-to-speech service.

Package deepseek provides DeepSeek's OpenAI-compatible LLM service.

Package elevenlabs is a streaming text-to-speech service backed by the ElevenLabs HTTP streaming API.

Package fireworks provides Fireworks AI's OpenAI-compatible LLM service.

Package fish is a streaming text-to-speech service backed by Fish Audio.

Package geminilive is a speech-to-speech service built on Google's Gemini Live API (BidiGenerateContent).

Package gladia is a streaming speech-to-text service backed by Gladia's Live STT v2 API.

Package google is a streaming LLM service backed by Google's Gemini API (generateContent with SSE).

Package groq provides Groq's OpenAI-compatible LLM and Whisper STT services.

Package hume is a streaming text-to-speech service backed by Hume AI's Octave TTS.

Package kokoro is a text-to-speech provider for a local Kokoro-FastAPI server.

Package lmnt is a streaming text-to-speech service backed by LMNT's HTTP "speech bytes" endpoint, which streams raw PCM.

Package mem0 adds long-term memory to a jargo voice agent, backed by a mem0 server (https://github.com/mem0ai/mem0).

Package minimax is a streaming text-to-speech service backed by MiniMax's T2A API.

Package mistral provides Mistral AI's OpenAI-compatible LLM service.

Package nebius provides Nebius AI Studio's OpenAI-compatible LLM service.

Package novasonic is a speech-to-speech service built on Amazon Nova Sonic, served over Bedrock's bidirectional streaming API.

Package nvidia provides NVIDIA NIM's OpenAI-compatible LLM service.

Package ollama provides a client for a local Ollama server through its OpenAI-compatible LLM endpoint.

Package openai provides OpenAI's LLM, STT and TTS services, plus the OpenAI-compatible LLM base that other providers (Groq, Together, Fireworks and the rest) wrap with their own base URL, key and default model.

Package openairealtime is a speech-to-speech service built on OpenAI's Realtime API.

Package openrouter provides OpenRouter's OpenAI-compatible LLM service, a gateway to many models behind one API.

Package perplexity provides Perplexity's OpenAI-compatible LLM service.

Package piper is a text-to-speech provider for a local Piper HTTP server (python -m piper.http_server).

Package qwen provides Alibaba's Qwen LLM over the DashScope OpenAI-compatible endpoint.

Package rime is a streaming text-to-speech service backed by Rime's HTTP API.

Package sambanova provides SambaNova Cloud's OpenAI-compatible LLM service.

Package soniox provides Soniox real-time streaming speech-to-text over its WebSocket API.

Package speechmatics provides Speechmatics' real-time streaming speech-to-text over its WebSocket API.

Package together provides Together AI's OpenAI-compatible LLM service.

Package whispercpp is a speech-to-text provider for a local whisper.cpp server.

Package xai provides xAI's (Grok) OpenAI-compatible LLM service.