A developer created MemoryGatekeeper, a Python class that gives AI agents a persistent memory of past tool-call failures using Mem0's local vector store. The gatekeeper checks semantic similarity agai…
Faisal Iitian released MindCache, a Python SDK with an MCP server for long-term memory in LLMs, reporting a 64% average rubric pass rate on his BEAM evaluation versus about 53% for Mem0. The system us…
DeepMem, a drop-in AI memory layer compatible with the Mem0 API, claims 2× faster response times and 10× lower cost, with self-hosted options at $0 and cloud plans starting at $1.9/month compared to M…
AI agent memory has become a crowded corner of AI infrastructure funding in 2025 and 2026, with startups like Letta, Mem0, and Zep building persistent memory systems on vector databases and retrieval …
A developer built Mythicator, a CLI tool that syncs a single per-repo memory file across context files read by Claude Code, Codex, Gemini CLI, and Cursor, preventing AI coding tools from forgetting de…
A final-year project called ANAMNESIS, running to April 2027, addresses a gap in LLM memory systems: facts recorded correctly can later become false as the world changes, and no existing system foreca…
Mem0's memory conflict resolver, which uses LLM-driven text similarity to automatically add, update, or delete memories, can silently delete facts that are still needed when it misinterprets context-s…
E J Ziyad launched UML (Universal Memory Layer) on July 31 as a shared memory system for Claude, ChatGPT, coding agents, and custom apps, using a graph-based architecture to store and retrieve durable…
A developer built a personalized long-term memory RAG stack on Fedora 44 using Mem0, Qdrant, and Open WebUI, with a FastAPI proxy that injects relevant past facts into chat prompts and auto-extracts n…
A developer's RE-call memory system, which never calls an LLM, scores lower on temporal reasoning (0.408 vs. 0.567) than Mem0 on the BEAM benchmark, but avoids any LLM cost and keeps data in Postgres.…
A developer discovered that the widely-cited LOCOMO benchmark for AI memory excludes 22.5% of its questions—those testing whether a system can correctly say 'I don't know'—and instructs models to neve…
A developer explains that LLMs are stateless and that the intelligence of a long-running agent lives entirely in the harness that manages memory injection. The post introduces Smriti, an open-source t…
A developer proposes using typed memory categories with Mem0 to improve LLM agent memory management, addressing failures from treating all facts equally. By storing memories as 'user', 'feedback', 'pr…
A developer's analysis of AI memory deletion reveals that consumer AI assistants like ChatGPT and Claude allow users to delete individual memories through settings panels, but cannot prove the underly…
Researchers introduce MemHop, a multi-hop memory benchmark of 1,000 questions at hop depths 1-5 across 10 social-network scenarios, and ProGraph, a two-layer memory architecture combining profile expa…
NVIDIA released Nemotron 3 Embed on July 16, and the 8B model immediately landed at #1 on RTEB with a score of 78.5%, topping Voyage 4 Large, OpenAI text-embedding-3-large, and Cohere embed-v4. The mo…
Chroma's July 2025 study of 18 frontier models found that performance degrades as input length grows, even before hitting hard context limits, and that models perform better on shuffled documents than…
Researchers propose MemTrace, a framework that transforms large language model memory pipelines into executable evolution graphs for fine-grained error tracing, and introduce MemTraceBench, a benchmar…
A developer testing a support-agent LLM with a memory layer found that the agent retrieved the correct fact (user on enterprise plan) but still responded with an incorrect suggestion to upgrade to ent…
A developer has created PVM (Persistent Vector Memory), an open-source Python library (~800 lines) that gives LLMs long-term memory without API keys or GPUs. It uses temporal weight decay, delta rule …