FSAR
中EN
Documentation · Core runtime
Documentation menu
Getting started
Project overviewPublic docs index
Config & development
Configuration referenceBuild, test, develop
Module reference
Modules
Core runtime
serverclicoretoolsprovidersmcp
Memory & data
memorylayoutnotificationsschedulerupdatesutils
Safety & execution
securitysandboxskills
Interface & extensions
frontendskinsocial

providers — LLM / ASR / TTS adapters

Language: 中文 | English · Back to module index

A uniform adapter-plus-dispatch structure for three modalities; presets ship in data/presets/.

Dir Description
llm/ presets.py, deepseek.py, google.py (Gemini native REST), thinking.py (thinking-effort mapping). Generic OpenAI-compatible goes through utils/llm_factory.py.
asr/ dispatch.py + adapters/ (faster_whisper, openai_compat, volcengine).
tts/ dispatch.py (selection/caching/retry) + cache.py (tts_cache.db) + adapters/ (azure, dashscope, edge, elevenlabs, minimax, openai_compat, volcengine).
pricing.py Model pricing tables and cost estimation (feeds the Usage page).
Source docs-public/modules/providers.mdView on GitHub ↗