🧠 External memory for AI assistants — facts, decisions, roles and history stored outside the context window. MCP server for OpenCode, Claude Code, Cline, Kimi Code.
-
Updated
Oct 6, 2026 - TypeScript
🧠 External memory for AI assistants — facts, decisions, roles and history stored outside the context window. MCP server for OpenCode, Claude Code, Cline, Kimi Code.
Symbiosis — A Manual for a Good Lineage (v2.0). A method for the human side of working with AI. EN/ES · CC BY-SA 4.0
ASCENT: scale-complementary external state for frozen long-context LLMs, evaluated on structured-fact retrieval (BABILong, RULER). Reproduction package for the manuscript under journal review: frozen artifacts, per-row outputs, SHA-256 manifests, CPU-only make reproduce.
Reference implementation and validation evidence for a resolution-first Context Router for AI external memory.
Research and training stack for AVA — a tool-using, memory-aware virtual assistant targeting 4 GB VRAM. Spans custom transformers, verifier-RL, external memory, multi-domain benchmarks, and Gemma 4 inference optimization.
Out-of-core CUDA bulk loading of Sort-Tile-Recursive R-trees for datasets larger than both VRAM and RAM, with a phase-identical CPU build as the control group.
An I/O-efficient implementation of (Binary) Decision Diagrams
七日之喵 · 7 天自主生命体实验 — 一个自主 AI Agent 的完整一生(D1-D7):研究论文(约 3.2 万字)+ 完整过程数据集 | CC BY 4.0
Real-time external memory and cognitive continuity layer for LLMs, autonomous agents, and model pretraining workflows. Built with zero PyTorch overhead using FastEmbed.
Shared Substrate — a discipline for sustained human–AI coupling on complex problems: layered external memory, verification oracles per abstraction layer, observability feedback loops, human-governed agent execution.
strange cross-platform, performance-oriented and interactive pangenome and large graph visualization
File-backed governance for Codex and Claude Code: bounded tasks, inspectable workers, validation evidence, and current project state.
Source-grounded research on graph world models for continual learning and harness-based agent steering.
External memory architecture for human-led Claude alignment and long-term AI workflow integrity (MIT)
Save any YouTube video with Shift+Y, then search every transcript locally to find the exact moment by any phrase you remember.
Controlled recall for fragmented experiments, evolving hypotheses, and high-information next steps.
RUMA: routed updatable memory architecture for versioned document updates
External memory layer for multi-agent collaboration
External Memory (Binary) Decision Diagrams
Persistent External Brain for AI Coding Agents. 6 memory tiers, sub-100ms recall, 107K+ memories, 21 MCP tools, 85+ REST endpoints. Built in Rust.
To associate your repository with the external-memory topic, visit your repo's landing page and select "manage topics."