Hey Product Hunt! 👋 While building RAG pipelines and autonomous AI agents in India, I kept hitting three walls with US-hosted vector databases: 1. 200ms+ round-trip latency to us-east-1 on every vector query. 2. Free tier pods pausing after a few days of inactivity. 3. Having to wire up external embedding APIs, BM25 + dense vector hybrid search, and TTL jobs for agent memory manually. That's why I built Iceberg DB (https://icebergdb.in) — a serverless vector database and 4-tier AI agent memory engine hosted in Mumbai (ap-south-1) with sub-20ms latency. What you get out of the box: ⚡ Sub-20ms Query Latency in India (DPDP Act 2023 compliant) 🔍 Built-in Hybrid Search (Dense + BM25 Reciprocal Rank Fusion) 🧠 Native 4-Tier AI Agent Memory (short_term, episodic, long_term, semantic) 📄 Built-in Neural Embeddings & direct PDF ingestion 📦 Official SDKs for Python (pip install iceberg-sdk), JS/TS (npm install iceberg-js), Rust, Go & .NET 🎁 Forever Free Tier (100,000 vectors, never pauses, no credit card required) I'd love for you to test it live at https://icebergdb.in and share your feedback!

Iceberg Vector Database

Serverless Vector Database & 4-Tier AI Agent Memory Engine

Serverless Vector Database & 4-Tier AI Agent Memory Engine