Letta vs Zep

Letta is freemium — a usable free tier with paid plans above it. Zep is freemium — a usable free tier with paid plans above it. Both are listed under Agent Memory & RAG-as-a-Service, so this is a like-for-like comparison. Neither is ranked above the other — Flocci carries no sponsored placement.

Letta vs Zep — straight answers

Letta vs Zep: what is the difference?

Letta is freemium — a usable free tier with paid plans above it and is listed for free BYOK plan supporting up to 3 stateful agents. Zep is freemium — a usable free tier with paid plans above it and is listed for free plan: 10,000 monthly credits, 2 projects, 200ms retrieval. Both sit in Agent Memory & RAG-as-a-Service.

Is Letta or Zep cheaper to start with?

Neither — Letta and Zep are both freemium, so the choice comes down to capability rather than cost. Both entries list what they uniquely do above.

Which should I choose, Letta or Zep?

Choose Letta if you need free BYOK plan supporting up to 3 stateful agents; choose Zep if you need free plan: 10,000 monthly credits, 2 projects, 200ms retrieval. Flocci AI Tools does not rank one above the other — it shows both feature sets side by side and lets the requirement decide.

Letta compared with Zep: pricing tier, category, listed capabilities and links.
 LettaZep
Pricing tierFree tier + paid plansFree tier + paid plans
Free to startYesYes
CategoryAI Agents & AutomationAI Agents & Automation
TypeAgent Memory & RAG-as-a-ServiceAgent Memory & RAG-as-a-Service
Listed capabilities
  • Free BYOK plan supporting up to 3 stateful agents
  • Stateful-agent runtime with persistent memory, desktop app, CLI and API
  • Open-source framework (formerly MemGPT) plus a managed cloud
  • Free plan: 10,000 monthly credits, 2 projects, 200ms retrieval
  • Graph RAG with automated context assembly from chat + business data
  • Temporal knowledge graph for tracking how facts change over time
Tagsstateful agents, agent memory framework, letta cloud, mcp memory server, open source agents, freeagent memory, graph rag, context engineering, chat memory, temporal knowledge graph, free
Websiteletta.comgetzep.com
Full pageLetta details →Zep details →
AlternativesLetta alternatives →Zep alternatives →

Also worth comparing

All Agent Memory & RAG-as-a-Service →

MemOS

Agent Memory & RAG-as-a-Service
freeNew
  • Memory operating system for LLM agents with hybrid retrieval and persistent memory
  • Open source and self-hostable, with a cloud option
  • Built for multi-agent and personal assistants
  • Backed by MemTensor, a Shanghai team
Visit

RAGFlow

Agent Memory & RAG-as-a-Service
freeNew
  • Open-source RAG engine with deep document parsing and agent workflows
  • Grounded citations and template-based chunking
  • Self-host with Docker; bring your own LLM
Visit

Basic Memory

Agent Memory & RAG-as-a-Service
freeNew
  • MCP server that stores AI conversation memory as local Markdown notes
  • Works with Claude Desktop and other MCP clients, and is readable in Obsidian
  • Everything stays on your disk
  • Open source core is free; optional paid cloud sync
Visit

Cognee

Agent Memory & RAG-as-a-Service
freeNew
  • Open-source memory platform that turns data into a knowledge graph plus vectors for agents
  • Python API with pluggable graph, vector and relational stores
  • Self-host free; cloud option exists
Visit

Graphiti

Agent Memory & RAG-as-a-Service
freeNew
  • Framework for building real-time temporal knowledge graphs for AI agents
  • Powers Zep's memory layer; runs standalone
  • Open source, works with Neo4j and other graph DBs
Visit

LightRAG

Agent Memory & RAG-as-a-Service
freeNew
  • Simple, fast graph-based RAG from HKU Data Science lab (EMNLP 2025)
  • Combines knowledge-graph and vector retrieval
  • Works with local or hosted LLMs
Visit