A persistent memory layer that works across Claude, Codex CLI, and Copilot sessions. You store facts as entityType/entityId/key triples in Postgres, then retrieve them exactly or via hybrid search when keys are fuzzy. It includes four Staff agents: Librarian handles writes and conflict resolution, Attendant injects context per turn and suggests when to skip redundant tool calls, Archivist decays stale facts in the background, and Resolutionist escalates conflicts you need to resolve manually. Setup is `npm install -g iranti && iranti setup`, then run `iranti claude-setup` or the equivalent for your tool. Reach for this when you need agents to remember deployment status, file changes, or session checkpoints across context resets without re-explaining every time.
DATABASE_URL*PostgreSQL connection string for the Iranti database.
IRANTI_API_KEYsecretAPI key for a remote Iranti instance (e.g. iranti.cloud).
IRANTI_URLBase URL of the Iranti API server (e.g. https://<slug>.iranti.cloud).
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