What it does

Deliver a concise peepshow summary to a Letta agent as one user-role message. Letta's MemGPT-style memory pipeline distils it into recall memory so the agent can later answer questions about the clip. Works with Letta Cloud and self-hosted deployments.

When to reach for it

Let a Letta agent remember every video shared with it, durably, across sessions
Self-hosted Letta: no token required — point `LETTA_BASE_URL` at localhost
Team knowledge agent: push support-team clips into a shared Letta agent inbox

Install

npm i -g peepshow

Use it

LETTA_BASE_URL=https://app.letta.com \
LETTA_AGENT_ID=agent_... \
LETTA_API_KEY=... \
peepshow ./repro.mov --sink letta

Make it automatic

peepshow sinks add letta
peepshow sinks add letta --when extension=mp4,mov
peepshow sinks add letta --when path=/Volumes/Work/

Configuration

LETTA_BASE_URL Base URL, e.g. `https://app.letta.com` or `http://localhost:8283`. required
LETTA_AGENT_ID Target agent id the message is posted to. required
LETTA_API_KEY Bearer token. Required for Letta Cloud; optional for self-hosted.

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 Letta sink automatically when three things are true:

the env vars below are exported in the agent's shell (or a project .env it can load),
the peepshow CLI is on PATH — install with npm 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 LETTA_BASE_URL="..."
export LETTA_AGENT_ID="..."

2. Register as an auto-sink

peepshow sinks add letta
peepshow sinks add letta --when extension=mp4,mov

3. Example LLM session

You → drop a .mov into Claude Code.
Claude → auto-invokes /peepshow:slides ./clip.mov. peepshow extracts frames + audio, the Letta sink 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'.

Letta — Post a run summary to a Letta (MemGPT) agent inbox.