Agentic Memory Vector

What it is

Agentic Memory Vector is a zero-dependency, in-memory vector database designed for AI agents. It supports Cosine Similarity and persistence, allowing AI applications to implement long-term memory without the need for complex infrastructure like Pinecone or Weaviate. The tool is implemented in TypeScript and uses Float32Array for efficient performance across Node.js, Edge runtimes, and browsers. With a small footprint and no external dependencies, it is suitable for lightweight applications where simplicity and speed are priorities.

Zero-dependency, in-memory vector database with Cosine Similarity support.

When it fits — and when it doesn’t

Known weaknesses

Each one links to the public issue, changelog or docs page it came from, where we have one.

1. 01

   Low adoption

   Only 3 weekly downloads and 1 GitHub star.

2. 02

   Limited features

   Focuses primarily on in-memory vector storage and Cosine Similarity.

What it does well

• Zero dependencies

  The npm package lists a dependency count of 0.

• Supports Cosine Similarity

  The description mentions support for Cosine Similarity.

• Lightweight

  The unpacked size is only 18,581 bytes.

Pricing

Free tier

Yes

Paid from

—

Model

Flat rate

Enterprise plan

No

Pricing last confirmed Oct 2, 2026.