which is the mapping between silero plugin version and silero vad version?

Thanks for reaching out, Filippo. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.

1.3.9 or 1.4.1

Thanks — you're asking which Silero VAD model/checkpoint is used by

livekit-plugins-silero

releases 1.3.9 and 1.4.1. I’m happy to clarify. Short answer: There is no published, explicit mapping that ties a

livekit-plugins-silero

release number (for example 1.3.9 or 1.4.1) to a specific Silero VAD model checkpoint/version. The Silero plugin integrates the open-source Silero VAD project and downloads the model files at runtime, but plugin docs and release notes do not list a one-to-one version mapping. If you need to know exactly which model each plugin version uses, here are practical ways to confirm it locally: 1. Inspect the plugin source for that release ◦ Check the plugin’s repo at the matching tag/release (for example in the livekit agents repo under livekit-plugins/livekit-plugins-silero) and look for the model download logic or any hardcoded model URL/checkpoint name. 1. Examine the downloaded model files used at runtime ◦ After installing the plugin version, run the model downloader (

uv run agent.py download-files

is the documented step to fetch model weights). ◦ Inspect the downloaded files (file names, any manifest, or checksum) in the plugin’s model cache folder — those often contain the checkpoint or version identifiers. 1. Check plugin release notes / changelog ◦ Look at the plugin release tag or changelog for that version; sometimes maintainers mention updates to model artifacts there. If you’d like, I can take one of the following next steps for you (please tell me which): • Look up the plugin source at the 1.3.9 and 1.4.1 tags and tell you what the model download references show. • Walk you step-by-step through checking the downloaded model files on your system (tell me your environment and where the agent is installed). • Provide tests to validate whether VAD behavior changed between the two plugin versions (sensitivity/false positive checks) if your goal is behavioral compatibility rather than a checkpoint name. Which option would you prefer? Sources: Silero VAD plugin | LiveKit Documentation | LiveKit turn detector plugin | LiveKit Documentation | Voice AI quickstart | LiveKit Documentation

2

Great — I’ll guide you through inspecting the downloaded model files so you can identify which Silero VAD checkpoint your installed plugin is using. First I need two quick details: • What OS/environment is the agent running on (Linux, macOS, Windows, Docker, or something else)? • Where is the agent installed or run from (path to the agent repo, or the container image name)? If you’re not sure, tell me how you start the agent (command you run). Once you share that, I’ll give step‑by‑step commands to: 1. Locate the plugin’s model cache/download folder. 2. List the downloaded model files and any manifest files that include a checkpoint name or URL. 3. Show how to inspect file contents or checksums that typically reveal the model version. If you'd like, paste the output of

ls -la

(or equivalent) for the agent directory after you run the download step (

uv run agent.py download-files

) and I’ll point out which files to check and how to interpret them.