A

ai-raccoon

Free

Give AI agents a memory that outlives the session: local-first SQLite per project, hybrid search, workspace sandboxes, and optional cloud sync.

Visit Website
Related by shared tags and keywordsnanobrowser1HelmMaximo AI
ai-toolsagent-memoryai-agentsdotnet-toolsembeddingaiai assistantautomation

Overview

AiRaccoon is an MCP server that provides AI agents with persistent, project-scoped memory and a searchable index of the project's own code. It runs locally on .NET 10 and SQLite, and includes hybrid keyword and semantic search, workspace sandboxes, a shared cross-project tier, and optional cloud sync. The server can be installed as a global .NET tool and integrated with MCP clients such as Claude Code, Hermes, or IDEs via JSON-RPC over stdio.

The system uses a proxy and an HTTP backend, with data stored in a memory.db file that combines FTS5 keyword search and vec0 vector search, fused with Reciprocal Rank Fusion. It supports a code corpus covering 37 source extensions across many programming languages, and can watch directories for changes to re-ingest content. Memory is partitioned by project id, and features include rating and decay, encryption at rest, and telemetry via OpenTelemetry.

AiRaccoon offers embedding models: a bundled default model (granite-embedding-small-english-r2) and options for downloaded Hugging Face models or remote OpenAI-compatible endpoints. Performance benchmarks on an Apple M4 show that the Neural Engine (coreml) option embeds twice as fast as the default WebGPU, with lower energy and CPU usage. The project is open source under the MIT license.

Pricing

Free

Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.

Visit Website

Prices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.

Use Cases

  • AI agents use AiRaccoon to maintain persistent, project-scoped memory across sessions.
  • Developers use AiRaccoon to search their codebase with hybrid keyword and semantic search.
  • Teams use workspace sandboxes to keep isolated in-progress notes that can be consolidated or discarded.
  • Users can enable optional cloud sync for S3 or Azure Blob.
  • Users can encrypt data at rest with a passphrase.
  • Users can monitor performance with OpenTelemetry metrics and traces.

Features

Stores persistent project-scoped agent memory

Searches project code separately from notes

Combines keyword and semantic search

Reingests watched files after changes

Isolates temporary workspace notes

Shares selected facts across projects

Expires memories that remain unused

Optionally syncs through cloud storage

Reviews

No reviews yet. Be the first to share your experience!

Rankings & collections