Solo-built Mnemos memory system recalls 2x more facts than Claude Code's native memory
RileyRalmuto · x · 2026-11-20
Developer Riley Ralmuto released the MCP and Hermes plugin for Mnemos, an agent memory system, along with previously withheld test results: across three tests run on two years of real work history, Mnemos recalled roughly twice as many facts as Claude Code's native memory (65% vs 35%, 41% vs 14%, 42% vs 22%), while being faster, cheaper and more context-efficient.
Key findings:
- Mnemos outperforms native memory across the board, delivering contextual memories before the agent goes digging through notes;
- Surprisingly, Mnemos alone beats combining Mnemos + Claude's native memory — stacking the two systems doesn't help.
Built by one person and their agents, it's now an official plugin in Nous Research's Hermes marketplace: pip install "mnemos-continuity[mcp, embeddings]".
More from coding & agent
- Self-hacking with abliterated models: biggest risks were infra side channels, not app code — cyb3rops · 2026-11-20
- Pocket City: a pixel-art Amsterdam canal simulation built entirely on real geodata — anselm · 2026-11-20
- Mnemos vs Claude Code memory test details: 65%/41%/42% recall across three blind tests — RileyRalmuto · 2026-11-20
- Kaggle and Eigent launch AgenticRL-Bench to test coding agents on real RL incidents — dejavucoder · 2026-11-20
- Suno MCP now lets you generate original songs inside ChatGPT and Claude — heyshrutimishra · 2026-11-20
- OpenAI Shows How to Connect Team Data and Custom MCP Servers to ChatGPT — OpenAI · 2026-11-20