Settings

Integrations and access docs for the transcript knowledge base.

CLI — `chromescribe`

A bash CLI that lets any project query, search, and pipe transcripts from this knowledge base. Useful for summarization pipelines, grep-across-videos, or feeding content into Claude/Pi from another project's CWD.

Install

Source lives at cli/chromescribe in the repo. Symlink it onto your PATH:

# one-time setup
chmod +x ~/projects/chromescribe/cli/chromescribe
ln -sf ~/projects/chromescribe/cli/chromescribe ~/.local/bin/chromescribe

# sanity check
chromescribe --help

Required env

The dashboard API is behind Cloudflare Zero Trust. The CLI auto-reads a service token from env — put these in ~/.bashrc (or ~/.profile for cross-shell support):

export CF_ACCESS_CLIENT_ID="<your-service-token-id>"
export CF_ACCESS_CLIENT_SECRET="<your-service-token-secret>"

Note: service tokens live in Cloudflare Zero Trust → Access → Service Auth. This page is itself behind Zero Trust, so only you see it.

Commands

Command	What it does
`chromescribe list [-n N] [--json]`	List newest transcripts (default 50)
`chromescribe search <q> [--json]`	Search titles, channels, video_id
`chromescribe get <id> [--json]`	Show one transcript's metadata
`chromescribe cat <id>`	Print full markdown to stdout
`chromescribe url <source-url>`	Look up a transcript by YouTube/TikTok URL
`chromescribe view <id>`	Print dashboard URL for this transcript

Recipes

Save a transcript into another project:

chromescribe cat 141 > notes/obsidian-vs-rag.md

Read it in your terminal:

chromescribe cat 141 | glow -     # markdown pager
chromescribe cat 141 | less

Summarize with Pi (or any LLM CLI):

chromescribe cat 141 | pi "summarize in 5 bullets"

Grep across everything about a topic:

chromescribe search "RAG" --json | jq -r '.[].id' | while read id; do
  echo "=== $id ==="
  chromescribe cat "$id" | grep -i "vector"
done

Open in browser:

xdg-open "$(chromescribe view 141)"

Env overrides

Var	Purpose
`CHROMESCRIBE_API`	Override API base URL (default: this worker)
`CF_ACCESS_CLIENT_ID`	Cloudflare Access service-token ID
`CF_ACCESS_CLIENT_SECRET`	Cloudflare Access service-token secret

HTTP API

All endpoints require CF Access service-token headers. Responses are JSON with { success, data } wrapper except /api/transcript/:id/content which is raw markdown.

Method & Path	Purpose
`GET /api/transcripts`	List all transcripts
`GET /api/transcripts/:id`	Metadata for one transcript
`GET /api/transcript/:id/content`	Raw markdown body
`GET /api/search?q=...`	Title/channel/video_id search
`GET /api/channels`	Monitored YouTube channels
`POST /api/transcribe`	JSON (URL + markdown) or multipart (audio)
`PUT /api/transcripts/:id`	Update title or markdown
`DELETE /api/transcripts/:id`	Delete transcript + R2 content

Settings

CLI — chromescribe

Install

Required env

Commands

Recipes

Env overrides

HTTP API

CLI — `chromescribe`