What problem does it solve?
Lightweight end-of-session memory review — scan the conversation just had for durable lessons and save them into Ralphy's tiered memory, so the next session starts already knowing. The chat-native analog of hermes-agent's background self-review: signals are user corrections, frustration, durable preferences, model/provider facts discovered the hard way, and non-trivial fixes.
Core Features & Use Cases
- Scan the conversation for signals: user corrections, frustration markers, durable preferences, discovered facts, and techniques that worked.
- Filter out non-actionable or ephemeral data to keep memory focused and useful.
- Write memory notes to the tiered store and surface them in chat via "saved to memory" prompts; support "forget" instantly to retire notes.
Quick Start
End a session and run memory-review to save durable lessons into memory.