Speaches

Speaches is an OpenAI API-compatible server that provides local speech-to-text and text-to-speech capabilities through faster-whisper, piper, and Kokoro models.

Speaches is an API server designed to bring local speech recognition and generation to existing OpenAI-compatible tools and SDKs. The software supports speech-to-text via the faster-whisper model and text-to-speech through the piper and Kokoro engines. Users can perform tasks like audio transcription, sentiment analysis, and async speech-to-speech interactions. The server loads and unloads models dynamically based on request specifications and periods of inactivity. It runs on both CPU and GPU hardware and supports deployment through Docker. Streaming transcription is available, allowing users to receive output via Server-Sent Events without waiting for the full audio to process.

Speaches at a glance

Best for: Software engineer
Runs on: Docker
Pricing: Free
Released: May 2024
Category: AI audio tool
Website: speaches.ai

Key features

OpenAI API compatible endpoints
Streaming transcription via Server-Sent Events
Dynamic model loading and unloading
Text-to-speech generation with piper and Kokoro
CPU and GPU hardware support
Async speech-to-speech interactions
Sentiment analysis on recordings
Docker-based deployment

Use it for

Generating spoken audio summaries from text
Performing sentiment analysis on audio recordings
Building real-time speech-to-speech applications
Adding local transcription to OpenAI-compatible tools

Companions

Tools that articles about Speaches recommend using with it, each with its job.

Ollama — Built to work with Ollama.
Open WebUI — Run AI on your own terms with complete data privacy and self-hosting control.
AnythingLLM — Runs fully on-device with no accounts, API keys, or token limits required.
LangChain — Framework for custom agents with memory and tools on local models.
Chroma — Local vector database for embeddings and semantic search.
Cline — Open source, any model, with approval on every step

Alternatives to Speaches

AssemblyAI — Industry-leading speech models deliver high accuracy and low latency for voice applications.
Piper — Runs locally and offers fast neural text-to-speech synthesis.
WeNet — Achieves state-of-the-art results on public speech datasets.
PaddleSpeech — It offers state-of-the-art models and streaming systems for speech tasks.
Kokoro-FastAPI — Generate hours of high quality speech with an OpenAI-compatible endpoint.
Speech-to-Speech — Fully modular, swappable components with local or hosted execution options.
FunASR — It supports offline, streaming, and edge deployment with multiple models.
Pipecat — Builds real-time voice and multimodal agents with frame-level control.
Vosk — It enables offline, low-latency speech recognition on small devices.
sherpa-onnx — Runs complex audio AI tasks offline on local devices and embedded hardware.
Willow — Fast, private, open source voice control using local hardware.
kokoro-onnx — Runs local TTS inference with efficient ONNX runtime performance.

All Speaches alternatives