peepshow/sinks
The peepshow sink catalogue
Nineteen places avideocanland.
Every peepshow run can fan out to any number of destinations at once. Pick what fits — databases, vector stores, object storage, chat, trackers, knowledge bases, the IDE. Each one is pure transform + thin CLI, fully tested.
SQLite
One file, zero server, full SQL. Ships with auto-schema and upsert.
Local SQL archiveReel #02 · databasePostgres
Write every run + frame into Postgres. Idempotent upserts, transactional, cloud-friendly.
Team SQL archiveReel #03 · storageS3-compatible storage
AWS · MinIO · Cloudflare R2 · DigitalOcean Spaces · Backblaze B2 · Wasabi · GCS.
Object storageReel #04 · integrationWebhook
The generic sink. If something speaks HTTP, this connects peepshow to it.
Generic HTTPReel #05 · chatSlack
Incoming-webhook push. Title, thumbnail, deep link to the first frame.
Team chatReel #06 · chatDiscord
Discord webhook. Thumbnails in the embed, full frames as attachments.
Community chatReel #07 · integrationGraphQL
Typed ingestion into Hasura · Postgraphile · Apollo · Supabase · Contentful · anything.
Typed HTTPReel #08 · knowledgeNotion
Creates a page with metadata properties, embedded thumbnails, and the full frame gallery.
Knowledge baseReel #09 · knowledgeObsidian
File a note, embed every frame, backlink from an index, all inside a local vault.
Local knowledgeReel #10 · integrationIDE attachments
Drop frames directly into Cursor · Windsurf · Zed · VS Code / Copilot attachments.
Editor integrationReel #11 · trackerLinear
GraphQL + uploads. One issue per run, with frames inline, team/label configurable.
Issue trackerReel #12 · trackerGitHub Issues
REST v3. Repo / labels / assignees configurable. Frames attached via image URLs.
Issue trackerReel #13 · observabilitySentry
Frames as attachments, metadata as tags, peepshow run ID as fingerprint.
ObservabilityReel #14 · vectorChroma
Text embeddings over the metadata + caption. Frame paths as IDs.
Vector DBReel #15 · vectorQdrant
Named vectors, payload metadata, filter-ready. Works with Qdrant Cloud or self-hosted.
Vector DBReel #16 · vectorPinecone
Serverless. Auto-detects namespace. Embedding of metadata + caption per frame.
Vector DBReel #17 · vectorpgvector
Keep runs + frames + embeddings in one Postgres. One database, full SQL over vectors.
Vector-in-SQLReel #18 · databaseMongoDB
One document per run, frames as a subdocument array. Works with Atlas or self-hosted.
Document storeReel #19 · knowledgeMemPalace
Per-run markdown + frames into a directory MemPalace mines, optionally auto-mined on every run.
Local AI memory