memnest v0.2.1

Full Changelog: v0.2.0...v0.2.1

memnest v0.2.0

Full Changelog: v0.1.0...v0.2.0

memnest v0.1.0

First public release of memnest, a local persistent-memory service for AI coding agents.

Highlights

• Local hybrid retrieval with multilingual e5 embeddings, Tantivy BM25, HNSW, RRF, and MMR.
• Five canonical memory tools over Streamable HTTP MCP, stdio MCP, and the JSON HTTP API.
• Workspace-isolated recall with a built-in cross-project playbook collection.
• Prompt-time recall through memnest hook and redacted transcript capture through memnest watch.
• SQLite-backed durable writes, transactional index repair, atomic corrections, and an exclusive writer lock.
• Separate AES-256-GCM vault with model-facing secret tools disabled by default.

Install on Linux

Supported release architectures are x86_64 and aarch64.

curl -fsSL https://raw.githubusercontent.com/Blue-B/memnest/v0.1.0/core/scripts/install.sh \
  -o /tmp/memnest-install.sh
bash /tmp/memnest-install.sh --user

The installer downloads the matching archive, verifies its SHA-256 checksum, installs the binary, registers a user systemd service, and checks /health.

Windows and WSL service-registration scripts are included, but this release does not publish a Windows binary.

Documentation

• README
• Operations
• Retrieval benchmarks
• Core changelog

Full Changelog: https://github.com/Blue-B/memnest/commits/v0.1.0