Normal package discovery
Install with OpenClaw’s plugin command. The bare package spec can resolve through ClawHub with npm as the package source.
Local TTS for OpenClaw
Give OpenClaw a local voice.
Install the native AverVOX plugin from ClawHub or npm and route OpenClaw speech through the local bridge—streaming Piper or Kokoro audio without a cloud speech API.
Published on npmListed through ClawHubStreaming PCMSource-visible
What it gives you
A native provider, not a shell-script afterthought.
The plugin integrates with OpenClaw’s provider system and uses the same capability-aware bridge contract as other AverVOX clients.
Streaming and cancellation
Audio can begin as soon as the first sentence is synthesized, and interrupted requests terminate their local work cleanly.
Warm model reuse
Keep Piper or Kokoro loaded in the AverVOX daemon and avoid paying process startup plus model load for every agent reply.
Setup
Install, point, verify.
Use the OpenClaw package manager for the plugin and the AverVOX integration installer for a cautious, self-checking configuration path.
Install AverVOX
Use OSS or Pro on the Linux machine that will synthesize OpenClaw responses. Confirm avrvx --speak works first.
Install the OpenClaw plugin
Run the normal plugin installation command for @avervox/openclaw-plugin, or follow the package’s pinned-version instructions.
Verify real audio
Run avrvx --install-integration openclaw when appropriate. It checks command availability and produces an actual audio sample rather than only validating YAML or JSON syntax.
Security disclosure
Treat native plugins as part of the agent’s trust boundary.
The safest marketing claim is the inspectable one: know where the package comes from, what it executes, what files it creates, and how it reaches AverVOX.
- Plugin source and package manifest are public
- Speech text is passed through stdin or protected local files
- Temporary directories are owner-only
- Child process groups are terminated on cancellation
- No cloud TTS credential is required
Common questions
What to know before you start.
Is the plugin published?
Yes. The package is published on npm and listed on ClawHub. This page links to npm, the ClawHub listing, and the public source repository.
Which OpenClaw versions were tested?
The current integration documentation records testing against OpenClaw 2026.6.33 from npm and 2026.7.2 from source. Check the package README for the newest compatibility statement before upgrading either side.
Does the plugin run inside OpenClaw?
Yes, as a native plugin. Native plugins are trusted code inside the host process, so review the source, install from the expected package, and pin versions in sensitive environments.
Does it support speech-to-text?
The current OpenClaw package is positioned primarily as a TTS provider. The underlying AverVOX bridge also supports STT for hosts that expose a compatible input-provider interface.
Can it use Kokoro?
When AverVOX Pro and Kokoro are installed, capability discovery can expose the available engine and voices. Piper remains the lightweight engine in both OSS and Pro.
Start with the workflow
Let OpenClaw speak locally.
Install the published plugin, verify it against your OpenClaw build, and use the warm bridge when low first-audio latency matters.