Frameworks

Letta (MemGPT)

Agents with persistent, self-editing memory — grown out of the MemGPT research — that learn across sessions.

As of 1 October 2026, Letta (MemGPT) ranks 19th of 34 in Frameworks on the Agent Reality Index — scored 47.5 of 100, led by market demand.

#1947.5Index / 100

Score breakdown

Market demand47
Developer adoption29
Real usage75
Momentum32
Production readiness74
Community30

Why this rank: the index weights market demand most heavily — how much Letta (MemGPT) is actually searched for and asked about inside AI tools.

Real-world demand

1.4kGoogle searches / mo
134AI / LLM searches / mo
1.4kProduct-site visits / mo
25kGitHub stars

Demand measured via Product-site traffic · confidence: high

GitHub repository ↗Full Frameworks board →Methodology →
Common questions

Letta (MemGPT) — frequently asked questions

What is Letta (MemGPT)?

Agents with persistent, self-editing memory — grown out of the MemGPT research — that learn across sessions. It is tracked in the Frameworks category of the Agent Reality Index.

What is Letta (MemGPT)'s Agent Reality Index ranking?

As of 1 October 2026, Letta (MemGPT) ranks 19th of 34 in Frameworks, with an overall Agent Reality Index score of 47.5 of 100.

How is Letta (MemGPT) scored?

The index blends six real-world signals — market demand (weighted most heavily), developer adoption, real usage, momentum, production-readiness, and community — into a 0–100 score, rather than relying on GitHub stars or vendor opinion.

Is Letta (MemGPT) open source?

Letta (MemGPT) is open source; its code is publicly available on GitHub.

How popular is Letta (MemGPT)?

Measured 1 October 2026, Letta (MemGPT) draws about 1.4k Google searches a month, around 134 AI and LLM searches, roughly 25k GitHub stars. Demand confidence is rated high.

Whichever tool you pick, the framework is the replaceable layer. Is your organisation ready?

Take the assessment →