MemGPT
PulseAugur coverage of MemGPT — every cluster mentioning MemGPT across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
Agent memory startups Mem0, Zep, Letta race to solve LLM forgetting problem · 1 source tracked
A new category of agent memory infrastructure is emerging, with companies like Mem0, Zep, and Letta vying for dominance. These solutions aim to solve the problem of Large Language Models forgetting information over long…
-
AI agents degrade due to prompt bloat, not context limits, experts say
A common failure mode in AI agents, even those using powerful models like GPT-5 or Claude, is a degradation in performance over time due to prompt bloat rather than context window limitations. This "context rot" leads t…
-
New Python library combats silent memory loss in LLM agents
A new Python library called memory-anchor has been developed to address the issue of context compaction silently degrading the memory of long-session LLM agents. This library snapshots the agent's state, including rules…
-
LLMs are stateless; agent intelligence resides in external harnesses
Large language models are inherently stateless, meaning they do not retain information between interactions. This fundamental characteristic necessitates that the intelligence and memory of AI agents reside not within t…
-
InferScale system optimizes LLM serving with reusable KV state
Researchers have developed InferScale, a new GPU-native system designed to enhance the serving of personalized large language models. InferScale addresses the issue of repeated prompt prefilling in memory systems by uti…
-
Mem0, Letta, Zep: AI Agent Memory Frameworks Compared
Three open-source AI agent memory frameworks—Mem0, Letta, and Zep—offer distinct approaches to memory management. Mem0 provides a universal CRUD API for vector embeddings, suitable for simple integrations into existing …
-
Letta Letta (formerly MemGPT) framework enables stateful AI agents
Letta Letta, previously known as MemGPT, is an open-source framework and server designed for creating stateful AI agents capable of long-term memory retention. This tool aims to enhance agent capabilities by providing a…
-
VEKTOR Slipstream beats GPT-4 on local memory benchmark
VEKTOR Slipstream, a local agent memory framework, achieved a 79% score on the LongMemEval benchmark, outperforming full-context GPT-4 by 12 points. This benchmark specifically tests real-world memory retrieval failures…
-
Eidetic OS leads AI memory tool comparison, stressing privacy
A comparison of AI memory tools highlights Eidetic OS, Mem0, Letta, Khoj, and Nucleus MCP, each offering different approaches to addressing LLM amnesia. Eidetic OS, developed by solo developer Paul Holland, emphasizes l…
-
AI agents struggle with true learning due to stateless models
Current AI agents, despite improvements like GitHub Copilot's session memory, fundamentally lack true learning capabilities because their underlying models are pure functions that reset with each query. While workaround…