Depths AI is building a unified storage layer called gyaan for AI Agents and applications. The aim of the package is to enable developers to integrate vector-embedding database, chat history, analytics logs, knowledge graphs and a memory agent layer on top it, within few lines of code.

Currently, gyaan is v0.1: The world’s cheapest S3-compatible vector embedding database.

The project is and will proudly remain open-source! For ease of usage, we encourage you to try out the free tier of the cloud version (currently in open-beta).

​
Brief Roadmap

Leveraging a unified storage core for S3-compatible storage, we would introduce the remaining aspects of storage layer as:

  1. v0.2: Chat history
  2. v0.3: Analytics logs
  3. v0.4: Knowledge graph
  4. v0.5: Memory agent layer
  5. v0.6-0.9: Improving storage flexibility, indexing quality, MCP support etc
  6. v1.0: Production ready!

​
Install Gyaan DB

We recommend using uv for managing your python projects. Command to add gyaan in an existing uv project is as follows

uv add gyaan

Or if you prefer pip installation for your python projects:

pip install gyaan

Run gyaan locally

Start using gyaan locally on your disk or connect it to your S3 bucket.

Access gyaan in the cloud

Integrate gyaan in your AI agent via the cloud client-side SDK.