Light
Dark
pro plan
Astronaut walking through rocky canyon landscape on planetary surface.

Pro $20 / month

For personal use
Everything in Free, plus:
Usage quota for open-weights models and Letta Auto
Pay-as-you-go pricing for additional models or overage
Up to 20 stateful agents
max lite plan
Spacecraft with landed on planetary surface with astronaut figure nearby.

Max Lite $100 / month

For professionals
Everything in Pro, plus:
Usage quota across all frontier model providers
5X higher limits on Letta Auto
Up to 50 stateful agents
Max plan
Colony base with dome structures and buildings on planetary surface.

Max $200 / month

For power users
Everything in Max Lite, plus:
Increased quota for frontier models
20X higher limits on Letta Auto
Early access to new features
API plan

API Plan $20 / month

For organizations and teams building their own applications on top of the Letta API. Unlimited agents with usage-based pricing.
Unlimited agents
$0.10 / active agent / mo
$0.00015 / sec tool execution
API key authentication
Pay-as-you-go LLM usage
How is tool execution charged?

Server-side tools on the Letta API incur a credit cost for CPU time required to execute the tool. Remote MCP tools are executed by the MCP tool provider, so do not have a credit cost. Letta built-in tools are executed for free. Client-side tools (e.g. bash tools executed by Letta Code) are run on the client not the server, so they do not incur any credit cost for execution.

How does the Max plan work?

The Letta Max plan is designed for agent power users that run token-intensive workloads, such as high-throughput agentic coding. The Max plan includes a generous request-based system with significantly higher usage limits than the Pro plan. Request quotas are reset multiple times per day.

Note that the Letta Max plan is for personal use only. Developers building applications on the Letta API for external end-users should use pay-as-you-go credits, in addition to included monthly credits in the Letta Pro plan.

What is Letta Code?

Letta Code is a memory-first coding harness, built on top of the Letta API. Instead of working in independent sessions, you work with a persisted agent that learns over time and is portable across models.

You can use the Letta Code CLI to run agents locally on your own computer, where they can access your local filesystem, and take other actions on your behalf. Learn more about Letta Code on our docs.

Where can I ask more questions?

Reach out to [email protected], or join our active community on Discord to chat with the Letta developer team and other Letta users!