Letta
Letta is an AI research lab shipping Letta Agent software focused on agents with persistent memory and ‘‘sleep‑time’’ background reasoning. The project publishes technical papers, docs, and an npm package; hosted services have a documented hosted-data policy distinguishing hosted versus self‑hosted data handling.
The linked official pages do not list plan prices.Is Letta right for you?
Good for
- Research or product teams experimenting with persistent-memory agents and background reasoning who want a research-forward agent platform and developer tooling.
- Research lab focus and productized agent software.
- Provides sleep-time agents (background reasoning) as a released feature.
Keep in mind
- Hosted endpoints retain message requests and responses (data used by Letta services).
- Letta may use collected hosted-service data to train future models.
- Buyers who need a public commercial pricing table or explicit hosted plan costs should confirm Letta’s pricing and deployment options with the vendor before committing.
Pricing and plan limits
The sources describe shipping software (an npm package) and hosted services but do not provide plan names, numeric prices, or billing commitments on the linked pages.
More about capabilities and limits
Documented strengths
Research lab focus and productized agent software.
Provides sleep-time agents (background reasoning) as a released feature.
Clear hosted-data collection boundary: hosted services collect message contents.
Limitations to consider
Hosted endpoints retain message requests and responses (data used by Letta services).
Letta may use collected hosted-service data to train future models.
Choosing a plan
Ask Letta for hosted versus self-host deployment details, pricing, and data-handling terms, and validate how sleep‑time agents integrate with your memory and storage systems.
Based on the linked product documentation.
Available on
Desktop, Api.
Official sources
Alternatives to Letta
Choose around the work you need to do.
Mastra
Use Mastra as an open-source TypeScript framework to build, observe, and deploy long-running, memory-enabled agents with observability and multi-provider model routing for production deployments.
TextCortex
Choose TextCortex to build and manage enterprise AI agents with a no-code visual builder, scheduled automations, and integrations into Slack, Teams, and browser workflows for context-aware tasks.
