Crate rustymcp 

Rusty Memory is an educational memory server that demonstrates how to wire semantic chunking, embeddings, and Qdrant together behind both an HTTP API and the Model Context Protocol. The crate exposes deliberate, well-commented modules so learners can trace the end-to-end data flow:

• config loads the runtime configuration and explains the required environment variables.
• processing contains the orchestration pipeline that chunks text, requests embeddings, and writes vectors to Qdrant while recording metrics.
• api and mcp surface the same processing primitives through REST and MCP tooling respectively.
• metrics, logging, and qdrant provide the supporting infrastructure.

The library can be embedded in other projects or used as documentation for students exploring how modern agent memory stacks are built in Rust.

Modules

api

HTTP routing and REST handlers. HTTP surface for Rusty Memory.

config

Environment-driven configuration management. Environment-driven configuration for Rusty Memory.

embedding

Embedding client abstraction and adapters. Embedding client abstraction and provider adapters.

logging

Structured logging and tracing setup. Tracing configuration and log routing.

mcp

Model Context Protocol server implementation. Model Context Protocol (MCP) integration for Rusty Memory.

metrics

Ingestion metrics helpers. Lightweight ingestion counters used for diagnostics.

processing

Document processing pipeline utilities. Document processing pipeline: chunking, embedding, and Qdrant orchestration.

qdrant

Qdrant vector store integration. Qdrant vector store integration.

summarization

Optional abstractive summarization client(s). Abstractions for generating abstractive summaries via local providers.