What problem does it solve?
Honcho centralizes cross-session user modeling and provides per-profile AI peer memory so Hermes can maintain consistent user preferences, isolate agent identities, and avoid repeated context reconfiguration across runs.
Core Features & Use Cases
- Persistent User Modeling: Builds a unified user representation from observed messages across sessions to remember preferences and facts.
- Per-Profile AI Peers: Assigns an AI peer to each Hermes profile so agents develop independent identities while sharing workspace-level user context.
- Observation & Recall Control: Fine-grained toggles for observeMe/observeOthers, recall modes (hybrid/context/tools), write frequency, and session mapping for cloud or self-hosted deployments.
- Operational Tooling: CLI commands and status checks for setup, sync, migration, and troubleshooting when memory is failing or peers are out of sync.
- Use Case: Configure Honcho for Hermes to remember a user's role and formatting preferences across multiple agent profiles while preventing one profile's in-progress conclusions from polluting another profile's reasoning.
Quick Start
Enable Honcho for the active Hermes profile using the interactive Honcho setup and then verify the connection and peer information with the honcho status command.