Axonio vs Letta.
A framework and platform for stateful agents, from the team behind MemGPT. Here is where each one fits, and where Axonio goes further.
| Capability | Axonio | Letta |
|---|---|---|
| What it is | Memory engine + hosted MCP server | Stateful agent framework + platform |
| Open-source core | MIT | Apache-2.0 |
| Self-hosting | Same engine as the cloud | Letta server |
| Vector retrieval | Hybrid vector + full-text + graph | Archival memory search |
| Knowledge graph | Cross-source, time-aware | Not a core primitive |
| Fork / diff / merge memory | Zero-copy fork, diff, merge, discard | Not documented |
| MCP server | Hosted + local, 44 tools | Agents can use MCP tools |
| Embedded / on-device | Offline, syncs when connected | Server-based |
| Importers from other stores | Mem0, Zep, Pinecone + 6 more | Not documented |
From Letta’s public documentation, last reviewed September 2026. Out of date? Tell us.
Choose Letta if
- You want an opinionated agent runtime where the agent manages its own memory blocks.
- You are building on the MemGPT research line and its agent development environment.
Choose Axonio if
- You already have an agent framework and need a memory backend, not a new runtime.
- You need to fork, diff and merge memory across agents or experiments.
- You need the same memory on robots and edge devices.
FAQ
Axonio or Letta?
Is Axonio a Letta alternative?
Yes, for agent memory. Letta is a framework and platform for stateful agents, from the team behind MemGPT. Axonio is an open-source memory engine with a hosted MCP server, a knowledge graph and Git-style branching of memory.
What does Axonio have that Letta doesn’t document?
Fork, diff and merge of an agent’s memory. You can test a policy or a risky plan on a branch and merge or discard it without touching production memory.
Other comparisons