Rusty Memory FAQs

Question 1

How does Rusty Memory handle document indexing and chunking?

Accepted Answer

Rusty Memory features automatic, model-aware token chunking to efficiently prepare documents for embedding. It then uses a deterministic embedding client to convert these chunks into vectors, which are stored in Qdrant for fast retrieval.

Question 2

How easy is it to set up Rusty Memory?

Accepted Answer

Rusty Memory is designed for a 'zero-drama setup.' The entire stack consists of a single Rust binary and Qdrant. You can get started quickly with minimal configuration, and a Docker Compose example is provided for local environments.

Question 3

Does Rusty Memory support local embedding models?

Accepted Answer

Yes, local-first embeddings via Ollama are supported by default, allowing you to use local models without cloud API keys or costs. It also offers flexibility to switch to hosted embedding models via environment variables.

Question 4

What interfaces does Rusty Memory offer for AI agents?

Accepted Answer

It provides a dual interface: both a standard HTTP API and an MCP (Machine-Coded Protocol) server. This allows for diverse client interactions and seamless integration with various AI agent frameworks like Codex CLI, Kilo, or Cline.

Question 5

What is Rusty Memory?

Accepted Answer

Rusty Memory is a minimal, local-first Rust vector-memory server built specifically for AI agents. It simplifies the process of storing and retrieving agent memory, integrating seamlessly with Qdrant for vector storage and Ollama for local embeddings.

Rusty Memory

Rusty Memory

Key Features

Use Cases

Key Features

Use Cases