audio.cppvsSpeaches
Two self-hosted projects, compared on the numbers selfhost.directory tracks live: score, popularity, release cadence and setup effort.
| audio.cpp | Speaches | |
|---|---|---|
| Score | 74/100 | 63/100 |
| Rank in category | #15 of 121 in AI | #1 of 3 in Audio |
| GitHub stars | 4 410 | 3 777 |
| Stars, last 30 days | +575 | +54 |
| Stable releases, 12 months | 12 | 0 |
| Latest release | v0.9.6 · 21 Nov 2026 | — |
| Setup | Advanced | Moderate |
| Official Docker image | Yes | Yes |
| License | Apache-2.0 | MIT |
| Main language | C++ | Python |
Which one should you pick?
audio.cpp scores 74/100 against 63/100 for Speaches, so it is the safer default.
audio.cpp ships more often: 12 stable releases in the past year, against 0 for Speaches.
If you want the quickest setup, Speaches is easier to run; audio.cpp needs more configuration.
audio.cpp is growing faster right now (+575 stars in 30 days).
Key differences
- License: audio.cpp is Apache-2.0, Speaches is MIT.
- Language: audio.cpp is written mainly in C++, Speaches in Python.
- Features only audio.cpp lists: GGUF Model Format, JSON Pipelines, Multi-backend Support, Music Generation and 6 more.
- Features only Speaches lists: Audio diarization, Dynamic model loading, GPU and CPU support, Gradio web UI and 8 more.
audio.cpp details →Speaches details →