# xmark

chat with your X (Twitter) bookmarks. xmark pulls your saved tweets, embeds them with OpenAI, and answers questions in plain english using Claude — with citations linking back to the original tweet.

website: https://xmark-sigma.vercel.app
docs: https://xmark-sigma.vercel.app/docs
llms-full.txt: https://xmark-sigma.vercel.app/llms-full.txt
openapi: https://xmark-sigma.vercel.app/openapi.json
status: https://xmark-sigma.vercel.app/api/health
help: https://xmark-sigma.vercel.app/help

## what xmark is

- a personal search-and-chat layer over your X bookmarks
- semantic retrieval (pgvector + OpenAI embeddings) over the full bookmark contents
- Claude (claude-sonnet-4-5) answers grounded in your bookmarks with [N] citations
- developer REST API at /api/v1 with scoped API keys + webhooks

## what xmark is not

- not a bookmark manager (read-only mirror of X bookmarks; remove unbookmarks on X too)
- not an X analytics dashboard (no engagement insights or follower metrics)
- not a public search engine (your bookmarks are private — RLS-gated per user)

## quick start

1. sign in with Google
2. connect your X account on /bookmarks (separate OAuth flow for read access to bookmarks)
3. click sync — xmark pulls and embeds your bookmarks
4. open /chat and ask questions in plain english

## pricing

$5/month single plan. unlimited sync, unlimited chat. no free tier.

## tech

- Next.js 15 + React 19 + Supabase Postgres (with pgvector)
- Twitter (X) OAuth 2.0 with PKCE for bookmark read access
- OpenAI text-embedding-3-small for semantic search (1536 dims)
- Anthropic claude-sonnet-4-5 for chat answers
- Stripe for billing