Open source code for ICLR 2026 Paper: Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions
-
Updated
Aug 20, 2026 - Python
Open source code for ICLR 2026 Paper: Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions
Cortex: A Bidirectionally Aligned Embodied Agent Framework for Long-horizon Manipulation
ATM-Bench: A benchmark for long-term personalized memory QA spanning ~4 years of multimodal data (images, videos, emails). Features referential queries, evidence-grounded answering, and multi-source reasoning. Paper: "According to Me: Long-Term Personalized Referential Memory QA"
Production-ready agent implementations: SimpleAgent, ReactAgent, MultiAgent, MemoryAgent, RAG variants, and more
Template for agentic Chat & Memory Agents with Dapr runtime, UV simplicity, and OpenAI Agents SDK.
A framework for resonance-based decision-making in artificial agents, combining celestial influences, memory feedback, and symbolic emergence.
[EMNLP 2026] MemoryCD: Benchmarking Long-Context User Memory of LLM Agents for Lifelong Cross-Domain Personalization
A production-ready LangGraph agent with long-term memory - built from scratch on Redis.
Claude Code MCP native extension delivering a zero-human closed loop: Auto Bug Retrieval → Intelligent Troubleshooting → Automated Structured Accumulation. Equips Claude Code with long-term troubleshooting memory and establishes a reusable bug knowledge library for individuals and teams.
🐘 Local-first causal knowledge graph for AI developer memory. Tracks decisions, rejected alternatives & reasoning — not just facts. MCP-native for Claude, Cursor & VS Code.
Control autonomous AI agents by enforcing behavior rules to prevent unauthorized actions, improve focus, and boost execution efficiency.
Vraksha remembers so you can focus on creating.
MemoryLoom is an MCP-compatible memory server for editors, IDEs, and AI clients. It provides persistent memory with a clean tool interface, metadata-aware retrieval, and pluggable storage backends.
Provide precise OpenClaw documentation and exact configs to your coding agent for accurate answers without guesswork or summaries.
Echolocation on a phone for blind indoor navigation: a biomimetic chirp, a from-scratch PyTorch CNN that classifies the space, and an adaptive Qwen Memory Agent that learns each user. Qwen Cloud Hackathon, MemoryAgent track.
Production Qwen MemoryAgent for Thai rice/sugarcane farmers on Alibaba Cloud. 4-signal hybrid recall (vector + keyword + recency + reinforcement), LLM supersession, IoT alerts, pump HITL, MCP. Live on KhawTECH NaLog.
Sovereign, persistent, worldview-aligned AI agent. Runs on your hardware. No cloud. No moral relativism. Pure truth hierarchy.
🧠 Enhance AI agents with long-term memory and self-evolution using this lightweight, high-performance knowledge base middleware built in Go.
To associate your repository with the memory-agent topic, visit your repo's landing page and select "manage topics."