Installation¶
This section covers how to install, set up, and upgrade Achew. Pick the install method that fits your setup, then proceed through first-run configuration.
Install methods¶
After installation¶
- First Run: Connect Audiobookshelf and (optionally) configure an LLM service.
- Upgrading: How to install the latest version of Achew.
- Storage and Backup: Where Achew keeps its data.
- Reverse Proxy: For remote access.
System requirements¶
Minimum¶
- 10 GB disk space
- 4 GB RAM
Recommended¶
- 8 GB+ RAM if you plan to use the larger transcription models.
- SSD. Highly recommended since many processes are I/O heavy.
- Apple Silicon Mac for faster transcription (enables MLX services via the native install).
Networking Requirements¶
- Achew must be able to reach your Audiobookshelf server.
- Cloud LLM providers (optional) used for AI Cleanup require internet access.
- Local transcription services require internet access for initial model downloads.
Environment variables¶
A few launch-time knobs can be set as environment variables:
| Variable | Default | Purpose |
|---|---|---|
HOST |
127.0.0.1 |
Bind address for the backend. Equivalent to --host (or --listen for 0.0.0.0). |
PORT |
8000 |
Listen port. Equivalent to --port. |
DEBUG |
false |
Enables debug logging and FastAPI auto-reload. Equivalent to --debug / --no-debug. |
ACHEW_WORKER_COUNT |
auto | Overrides the number of parallel workers used for audio analysis. Equivalent to --workers. See Tuning the worker count. |