Developer documentation

CommentHunter is built for agents

Every extraction workflow used by the dashboard is published as a REST capability, an OpenAPI document, and plain-text LLM files. Agents can extract comments, reviews, posts, followers, leads, transcripts, and OCR without scraping the UI.

REST API

Supabase Edge Functions with Bearer auth. Same operations as the extractor UI.

API reference →

OpenAPI 3.1

Machine-readable schema for codegen, Postman, and agent tool calling.

openapi.json →

LLM files

llms.txt index plus expanded llms-full.txt for ChatGPT, Claude, and Cursor.

llms.txt →

One capability contract

CommentHunter currently exposes extraction, billing, and account operations covering comments, reviews, universal product types, checkout, and subscription status. Each operation has a REST route, JSON request/response schema, credit class, and generated reference page.

New product workflows are documented here first so agent support cannot become an afterthought. The dashboard, OpenAPI document, and LLM files reuse that contract.

Start in two minutes

curl "https://jdstzpeukxcxxnodvess.supabase.co/functions/v1/extract-comments" \
  -H "Authorization: Bearer $COMMENTHUNTER_ANON_OR_USER_JWT" \
  -H "apikey: $COMMENTHUNTER_ANON_KEY" \
  -H "Content-Type: application/json" \
  -d '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"limit":100,"includeReplies":true}'

Then read Quickstart for auth, credits, and a full first export.

Security and limits

Authenticated user JWTs deduct credits from the signed-in account. Public URLs only — private posts return PRIVATE_OR_UNAVAILABLE. Credit multipliers apply per platform and product type. Rate limits and provider quotas still apply. Destructive account or billing writes should be confirmed with the user.