Fully offline Qwen3 TTS for your agent
PendingStatic analysis audit pending.
Overview
No static analysis result has been recorded yet. Pattern checks will appear here once the artifact has been analyzed.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If the upstream repository or its dependencies change or are compromised, installing the skill could run different code than the artifacts reviewed here.
First-time setup executes a remote installer from GitHub's mutable main branch rather than a pinned, reviewed artifact; this is especially notable because the registry reports no install spec.
bash <(curl -fsSL https://raw.githubusercontent.com/daMustermann/claw-qwen3-tts/main/install.sh)
Avoid automatic curl-bash installation. Prefer a pinned commit or packaged ClawHub artifact, inspect the installer before running it, and pin dependencies/model revisions where possible.
If configured, these tokens let the local server send audio messages through your messaging accounts.
The skill supports Telegram bot and WhatsApp Business API credentials, which can authorize sending messages from user-controlled accounts.
"telegram": { "bot_token": "", "default_chat_id": "" }, "whatsapp": { "phone_number_id": "", "access_token": "" }Only configure messaging credentials if needed, use least-privilege tokens where possible, keep config files private, and confirm recipients before sending.
Audio sent through these integrations leaves the local machine and is shared with the selected messaging provider and recipient.
The WhatsApp integration uploads local audio to Meta's Graph API; the Telegram sender similarly posts audio to Telegram's Bot API.
upload_url = f"{GRAPH_API_BASE}/{phone_number_id}/media" ... files={"file": (Path(ogg_path).name, f, "audio/ogg; codecs=opus")}Treat messaging as an online sharing action, not offline TTS. Send only user-approved generated audio and verify the destination account or phone number.
Cloned voice material may remain available to the agent after the original task unless you delete it.
Designed or cloned voice profiles are persistent local data that can be reused in later sessions.
Voices are stored locally in `voices/` and persist across sessions. The agent automatically prompts to save after every voice design or clone operation.
Save cloned voices only with consent, use clear names, and delete voice profiles you no longer want retained.
The TTS server may keep running after the immediate request, consuming resources and exposing its localhost API.
The server is intentionally launched as a background process and tracked by a PID file until stopped.
nohup python -m uvicorn tts_server:app ... > "$LOGFILE" 2>&1 & echo $! > "$PIDFILE"
Stop the server when finished using `scripts/stop_server.sh` and review `server.log` if behavior is unexpected.