What it does
[Redis](https://redis.io) 8 ships [RediSearch](https://redis.io/docs/latest/develop/interact/search-and-query/) as a first-class module — full-text + vector search over hashes. This sink writes one hash per peepshow run (`<index>:<runId>`) and issues `FT.CREATE … IF NOT EXISTS` on first run so a paired full-text/vector index is materialised lazily. It shells out to the system `redis-cli` (no Node client dependency, zero install for users who already run Redis) and passes the password via stdin's `AUTH` command so secrets never appear in `ps aux`. When the FT module isn't loaded the hash still gets written; the sink logs the skip and exits 0.
When to reach for it
- Microsecond-latency keyword + vector search over peepshow runs, served from a Redis you already operate
- Pair with Redis Streams / Pub/Sub fan-out — same datastore, same operational footprint
- Embed peepshow into an existing Redis-backed agent memory layer without standing up a second search service
Install
npm i -g peepshow
# Redis 8+ with RediSearch module (preferred) or redis-stack
brew install redis # or: https://redis.io/docs/latest/operate/oss_and_stack/install/Use it
REDIS_URL="redis://127.0.0.1:6379" \
peepshow ./demo.mp4 --sink redis-vectorMake it automatic
Register the sink once — every run fires it afterward. Scope by --when so it only runs for matching videos.
peepshow sinks add redis-vector
peepshow sinks add redis-vector --when extension=mp4,mov
peepshow sinks add redis-vector --when path=/Volumes/Work/Configuration
REDIS_URLRedis connection URL. Default `redis://127.0.0.1:6379`.REDIS_INDEXFT index name + hash prefix. Default `peepshow:runs`.REDIS_PASSWORDPassword sent via the `AUTH` command on stdin (kept out of argv).REDIS_BINOverride the `redis-cli` executable path. Default `redis-cli` (must be on PATH).PEEPSHOW_FRAME_BASE_URLWhen set, the first frame URL is written to the `thumbnail_url` hash field.
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 Redis Vector 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
This sink has no required env vars — it writes to a local path. Pass the destination via --sink-arg:
REDIS_URL="redis://127.0.0.1:6379" \
peepshow ./demo.mp4 --sink redis-vector2. Register as an auto-sink
peepshow sinks add redis-vector
peepshow sinks add redis-vector --when extension=mp4,mov3. Example LLM session
You → drop a
.movinto Claude Code.Claude → auto-invokes
/peepshow:slides ./clip.mov. peepshow extracts frames + audio, theRedis Vectorsink 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'.