OpenedAI-Speech

Fully offline

Drop-in OpenAI TTS-compatible server. Self-hosted, talks to local voice models.

Editorial verdict: “Best 'drop-in local TTS for OpenAI clients'. Bridge solution for existing pipelines.”

Voice / transcription

Free

AGPL-3.0

★ 4.2 / 5

GitHub ★ 1,500

↗ GitHub

Compatibility at a glance

Which runtime + OS combos this app works against. Source of truth for "will it run on my setup?"

§ Runtimes supported

openai-compat

§ OS / platform

linuxmacoswindows

§ Hardware + model hint

Minimum VRAM

4 GB

Recommended starter model

Piper voices or Coqui XTTS-v2

→ Build the rest of the stack with /stack-builder → Pick a GPU for this app

What it is

OpenedAI-Speech mimics OpenAI's TTS API endpoint, but routes to local voice models (Piper, XTTS, OpenVoice). Drop a config, point any OpenAI-TTS client at it, get local voice synthesis. Pairs well with assistants that already support OpenAI TTS — drop-in replacement.

✓ Strengths

+True drop-in for OpenAI TTS API
+Multiple voice backends (Piper fast, XTTS quality)
+Docker compose works first try

△ Caveats

−Setup leans technical (docker-compose, model paths)
−Smaller user community than Whisper-side tooling

About the Voice / transcription category

Transcription, speech-to-text, or text-to-speech.

§ Other voice / transcription apps

MacWhisper

Best Whisper desktop app on macOS. Pay once, transcribe locally forever.

Buzz

Best open-source Whisper desktop app. Cross-platform, free, less polish than MacWhisper.

Where to go from here

Stack Builder →

Pre-filled with this app's recommended use case + budget tier. Get the full rig + runtime + model picks.

Back to /apps →

The full directory — filter by category, runtime, OS, privacy posture, or VRAM.

Runtimes (/tools) →

What this app talks to: Ollama, vLLM, llama.cpp, MLX, LM Studio. The upstream layer.

Community benchmarks →

Did this app work for you on a specific rig? Submit the benchmark — it powers the model + hardware pages.