What it does
Push every frame into a Weaviate class as a single batch. Weaviate's vectorizer (`text2vec-openai`, `text2vec-cohere`, etc.) handles embedding server-side, and multi-tenancy is supported via `WEAVIATE_TENANT`.
When to reach for it
- Semantic video search with Weaviate's built-in vectorizer modules
- Multi-tenant SaaS where each customer owns a Weaviate tenant
- Mix peepshow frames with other indexed objects in the same class
Install
npm i -g peepshowUse it
WEAVIATE_URL=http://localhost:8080 \
peepshow ./scene.mp4 --sink weaviateMake it automatic
Register the sink once — every run fires it afterward. Scope by --when so it only runs for matching videos.
peepshow sinks add weaviate
peepshow sinks add weaviate --when extension=mp4,mov
peepshow sinks add weaviate --when path=/Volumes/Work/Configuration
WEAVIATE_URLBase URL of Weaviate. Trailing slashes stripped. requiredWEAVIATE_CLASSClass to insert into. Default `PeepshowFrame`. Must exist.WEAVIATE_API_KEYBearer API key for Weaviate Cloud / auth-enabled installs.WEAVIATE_TENANTTenant name for multi-tenancy.
Use with an LLM agent
Every peepshow sink reads its config from env vars and receives a single JSON payload on stdin. An LLM agent (Claude Code, Cursor, Windsurf, Gemini, Codex) can drive the Weaviate sink automatically when three things are true:
- the env vars below are exported in the agent's shell (or a project
.envit can load), - the
peepshowCLI is onPATH— install withnpm i -g peepshow, - a peepshow auto-sink is registered for the run (optional but recommended — makes invocation zero-argument).
1. Set the environment
# Add to ~/.zshrc, ~/.bashrc, or a project .env the agent can load
export WEAVIATE_URL="..."2. Register as an auto-sink
peepshow sinks add weaviate
peepshow sinks add weaviate --when extension=mp4,mov3. Example LLM session
You → drop a
.movinto Claude Code.Claude → auto-invokes
/peepshow:slides ./clip.mov. peepshow extracts frames + audio, theWeaviatesink forwards the run to the configured collection. Claude replies with a summary and a link to the created record.
The transcript text is a natural embedding target — index it next to the frames.
Write your own
A sink is any executable that reads the --emit json payload on stdin. Shell, Node, Python, Go — the spec's in docs/PLUGINS.md. Register persistent ones with peepshow sinks add-cmd 'your-command'.