engram-memory
★ 38
vs
letta
★ 25.2k
engram-memory vs letta
engram-memory: Engram Memory provides AI agents with persistent, self-hosted memory, eliminating session-end forgetfulness and API costs. It features a three-tiered recall engine for efficient storage, search, and retrieval of semantic embeddings on your own hardware.; letta: Letta is a powerful platform for building stateful AI agents equipped with advanced memory, enabling them to learn and self-improve over time. It provides both a command-line interface for local agent execution and a comprehensive API with Python and TypeScript SDKs for seamless integration into applications.
01
TL;DR
engram-memoryMemory & Context · ★ 38
Providing long-term memory for AI agents (e.g., Claude Code, Cursor) to remember codebase details, user preferences, and architectural decisions across sessions.
lettaMemory & Context · ★ 25.2k
When to choose:
- You want a long-lived agent that remembers past work and rewrites its own memory and context over time.
- You want one agent reachable from CLI, desktop, browser, mobile and messaging apps, with subagents, schedules and hooks.
- You want memory you can inspect: the README says context is tracked in git and can sync to your own repository.
When to skip:
- You only need to add memory to an agent you already use through MCP; a small MCP memory server is a lighter fit.
- You rely on older Letta tutorials or the V1 API server: the project restructured, and the V1 server is retired.
- You want everything to work without signing in: Letta Cloud is the default backend and some features need it.
02
Side-by-Side Comparison
Field
engram-memorylettaCategory
Memory & Context
Memory & Context
Stars
★ 38
★ 25.2k
License
NOASSERTION
Apache-2.0
Install Command
docker run -d --name engram-memory --restart unless-stopped -p 6333:6333 -p 11435:11435 -p 8585:8585 -v engram_data:/data engrammemory/engram-memory:latest
npm install -g @letta-ai/letta-code
Updated
4mo ago
5d ago
Open Source
Yes
Yes
Tags
AI Agent Memory, Self-hosted, Vector Database
Stateful AI, Agent Platform, Advanced Memory
03
Features
engram-memory01Three-tiered recall engine (Hot-Tier Cache, Multi-Head Hash Index, Hybrid Vector Search) for fast and efficient memory retrieval.
02Persistent memory management with semantic embedding (store, search, recall, forget, consolidate) for AI agents.
03Local-first operation with self-hosted components (Qdrant, FastEmbed, MCP server) for data privacy and zero API costs.
04Kuzu-backed entity graph for tracking entities, co-retrieval patterns, and spreading activation.
05Advanced memory capabilities including conflict detection, temporal queries, and an "answer from memory" function.
06Hive Memory for multi-device sharing, enabling shared memory pools across different agents and devices.
letta01Build stateful AI agents with advanced memory
02Agents capable of learning and self-improvement over time
03CLI tool for running agents locally in your terminal
04Comprehensive API for integrating agents into applications
05Support for skills and subagents; fully model-agnostic
04
Use Cases
engram-memory↳Providing long-term memory for AI agents (e.g., Claude Code, Cursor) to remember codebase details, user preferences, and architectural decisions across sessions.
↳Enabling multi-agent systems to share a common knowledge base and collaborate effectively through shared memory pools (Hive Memory).
↳Answering complex questions by synthesizing information from an agent's stored memories, both locally and optionally with cloud enhancement.
↳Auditing an agent's knowledge at specific points in time or over date ranges using temporal queries for incident analysis or historical context.
↳Storing and retrieving user preferences, facts, decisions, goals, and insights efficiently and semantically for personalized agent interactions.
letta↳Running AI agents locally in your terminal for coding and general tasks.
↳Integrating stateful AI agents into custom applications using the API and SDKs.
↳Developing self-improving AI systems that adapt and learn continually.
05
Best For
engram-memory—
lettaMost PopularEssentialTrending
FAQ
FAQ
What is the difference between engram-memory and letta?
Both engram-memory and letta are in the Memory & Context category. engram-memory has 38 stars, while letta has 25.2k stars.
Which is better, engram-memory or letta?
The best choice depends on your use case. Choose engram-memory if Providing long-term memory for AI agents (e.g., Claude Code, Cursor) to remember codebase details, user preferences, and architectural decisions across sessions., and letta if Running AI agents locally in your terminal for coding and general tasks..
Is engram-memory free or open source?
Yes, engram-memory is open source on GitHub (NOASSERTION).
Is letta free or open source?
Yes, letta is open source on GitHub (Apache-2.0).
Cloud Infrastructure
AdDeploy on DigitalOcean
Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.
→