A agentindex
Comparison

CrewAI vs Letta

A like-for-like, spec-level comparison. Both entries are verified against their docs and repo.

CrewAI

Role-playing multi-agent crews.

  • + Intuitive role-and-task mental model
  • + Fast to prototype multi-agent flows
  • − Less fine-grained control than LangGraph
  • − Can be token-inefficient at scale

Letta

Stateful agents with long-term memory (formerly MemGPT).

  • + First-class long-term memory and stateful agents
  • + Lets agents persist and self-edit context
  • − Narrower focus than general orchestration frameworks
  • − Memory model adds complexity
Spec CrewAI Letta
Role Agent framework Agent framework
Tags agent-framework, library, multi-agent agent-framework, library, multi-agent
License MIT Apache-2.0
Open source Yes Yes
Self-hostable Yes Yes
MCP support Yes Yes
Pricing freemium freemium
Starting price BYOK usage cost Self-host free
Models claude, gpt, gemini, byok claude, gpt, gemini, byok
Languages python python
GitHub stars 54.4k 23.5k
Last activity 2026-06-27 2026-06-26