AI tools
BrainMCP
A semantic long-term memory MCP server for LLMs.
A Model Context Protocol server in Go that gives any LLM persistent semantic memory: vector embeddings via Gemini, contexts to organise topics, tagging, client collaboration and an LLM-assisted synthesis tool that answers questions from the stored memories. Backed by chromem-go for a lightweight local vector DB.
Specifications
- Semantic search via Gemini embedding-001 (768-d Matryoshka)
- Named contexts · tags · multi-client collaboration
- Persistent local store (binary + JSON), graceful shutdown
- Task-specific embeddings (RETRIEVAL_DOCUMENT vs RETRIEVAL_QUERY)
- Built-in interactive test CLI · works with any MCP client
More from AI tools
-
DSPLabAn AI-assisted lab for audio DSP.
A development environment and diagnostic laboratory for DSP engineering — Vult source on the left, an isolated AudioWorklet engine on the right, and an autonomous AI agent that can compile, probe, sweep and verify your DSP without leaving the browser.
-
AI GatewaySelf-hosted, multi-provider LLM gateway.
A lightweight Go gateway that sits between your apps and any LLM provider. Each client gets its own API key with independent backend, model whitelist, rate limits and quotas — all OpenAI-compatible.
-
Meeting AssistantPrivacy-first meeting transcription & analysis.
A C++17 tool that turns spoken meetings into structured, actionable knowledge — local Whisper transcription that never leaves your disk, deep AI analysis with role-specific personas, Obsidian + GitHub/GitLab integration and corporate-grade HTML reports with mind maps.
-
translator-aiMulti-provider AI translator for JSON i18n files.
A fast i18n CLI on npm for translating JSON locale files using Google Gemini, OpenAI or local Ollama / DeepSeek. Incremental cache only re-translates changed strings, multi-file deduplication kills duplicate API calls, and an MCP server lets LLM-driven workflows hand off translations programmatically.