A local-first memory system that stores agent decisions, lessons, and context in SQLite with semantic search, FTS5, and relationship tracking. Runs on your machine with no cloud dependency. Core operations include capturing typed memories with timestamps and contradictions, retrieving ranked context slices (semantic + full-text hybrid), and automated consolidation to prevent noise buildup. Pro adds multi-agent coordination with file claims and session rosters, LLM routing across providers, and knowledge base indexing for PDFs and docs. Tested at 95.4% on LongMemEval. Reach for this when you need session context to persist and compound locally instead of dumping markdown files or full chat history into every prompt. Built for agents that need institutional memory without the token overhead.
io.github.ericm1018/skillfm-llm-cost-optimizer-openai-anthropic-usage
io.github.mikerawsonnz/llm-orchestration-agent
io.github.mikerawsonnz/authenticated-llm-agent
labforgedev/copilot-memory-mcp
csoai-org/agent-prompt-injection-firewall-mcp
io.github.mikerawsonnz/authenticated-multi-llm-agent