v0.9.0 · MIT License
Stores what AI agents learn — thoughts, decisions, session events and repo facts — in PostgreSQL with pgvector and serves it back through versioned MCP tools with hybrid vector-plus-full-text retrieval and budgeted context packs. Enables multiple agents to share one scoped, role-isolated long-term memory, with deliberate promotion of distilled candidates into durable memories.
v1.0.0 · MIT License
Enables AI agents to run Meta (Facebook and Instagram) advertising: pull spend, CPA, ROAS and breakdown reports, audit account health, build and boost campaigns, ad sets and ads, and pause or adjust budgets, all starting paused until the user approves going live. Also supports brand memory, change logs and an optional daily spend cap.
v1.0.1 · AGPL 3.0
Enables local MCP clients to expose and execute typed tools through a bounded native runtime with Ed25519 admission, per-session budgets, rate limits, schemas, and cached outputs.
v0.1.0 · MIT License
Enables agents to run neural web searches, gather token-budgeted code context and documentation excerpts, and crawl individual pages through the Exa API.
— · Apache 2.0
A local Python + SQLite MCP memory service that stores persona, shared experiences, facts, agreements, and unfinished work, then restores them within a context budget so a new model can continue the same conversation. It supports offline original-text persistence, Chinese keyword retrieval, provenance-tracked memories, persona versioning/rollback, and progressive fading of older events.
— · MIT License
Enables Claude Code and Codex to write, update, and retrieve durable long-term memories that survive session boundaries and context compaction. All recall runs locally against a single SQLite file using hybrid BM25, embedding, and recency ranking cut to an explicit token budget, with no cloud dependency.