LetX
Real-time Collaborative LaTeX Editor
productLive product04 — The archivist / product / Shahriar LabsPrivate source
Context-Heavy.
Knowledge-Graph API for AI Agents
Public product website · 202601 / The idea
Context-Heavy is a multi-tenant knowledge-graph API: store, traverse and query context about codebases, people and systems with pgvector.
Context-Heavy is a multi-tenant REST API that stores, traverses and queries context for AI agents as a knowledge graph. Backed by PostgreSQL + pgvector for semantic search, Redis for caching, recursive CTE graph traversal, JWT and Google OAuth for auth, and Terraform-managed AWS ECS Fargate deployment.
02 / What matters
- 01
Multi-tenant REST API for AI agent memory
- 02
pgvector semantic search + recursive-CTE graph traversal
- 03
JWT + Google OAuth authentication
- 04
Production AWS deployment via Terraform
71.2%Recall@10 · 304-question slice
60.1%nDCG@10 · same slice
Evidence / Scope
71.2% recall@10 and 60.1% nDCG@10 on a 304-question, two-conversation LoCoMo slice. Retrieval-only evaluation; this is not a full LoCoMo answer-quality result.
Shahriar Labs research notes ↗Next / Keep exploring
There’s more.
QuantumSketch
AI Educational Video Generator
productLive productComiKola
AI Comic & Webtoon Platform
productLive productHave a challenge in this space?
Let’s talk about it ↗