/opt/autclaw, so recreating the container with a different image does not lose it.
The two images
The ffmpeg in the full image handles audio and video transcoding:
- Feishu: voice to Opus, video to MP4
- WeChat Official Account: voice to AMR, AMR to WAV
- Qianxun: voice to AMR, video to MP4
View logs
Stop or remove the container
/opt/autclaw.
Update the image
Switch between slim and full
Only the image name changes; keep the same volume::latest to return to the slim image.
Add ffmpeg temporarily
If you want to try transcoding on the slim image, install ffmpeg inside the running container. It takes effect immediately:Web screenshots (Chromium)
Chromium is about 300MB, and the full dependency tree usually pushes it past 400MB, so neither image bundles it. Install it when needed:fonts-wqy-zenhei is about 16MB and covers common characters; fonts-noto-cjk is about 91MB and has broader coverage.
The container runs as root by default, so Chromium must be started with --no-sandbox. Without it you will see Running as root without --no-sandbox is not supported; add the flag in your plugin (chromedp: chromedp.Flag("no-sandbox", true)).
Long-term option: build your own image
Bake the dependencies into the image so they survive container recreation. If you only need transcoding, use:full directly and skip the build. If you need transcoding plus screenshots, build your own image:
Use Docker Compose
Save the following ascompose.yaml, then run docker compose up -d:
- Change variants: set
imagetonolan1121953/autclaw:full. - Change directories: replace
/opt/autclawon the left side ofvolumeswith your NAS path. - Change ports: replace the host port on the left side of
"8200:8200".
Related docs
- First-time install: Deploy autClaw
- Persistent paths: Runtime directories and paths
- In-app upgrades: How to upgrade autClaw