14-stage Fusion Pipeline for LLM token compression — reversible compression, AST-aware code analysis, intelligent content routing. Zero LLM inference cost. MIT licensed.

Auditable context engineering for AI agents: context optimization, recoverable context compression, receipts, answer verification, and MCP for Claude Code, Codex, OpenClaw.

Portable CC-inspired skills for memory, verification, multi-agent coordination, context compression, and proactive coding-agent workflows.

A unified toolkit for efficient and effective coding agents (Karpathy principles, Caveman, Ponytail, RTK, CodeGraph, Context-Mode). Minimal setup under 30 seconds. Any OS.

Official Implementation for the ACL 2026 paper "SARA: Selective and Adaptive Retrieval-augmented Generation with Context Compression"

A drop-in proxy that compresses bloated code context in real-time, cutting LLM API costs by 50–80% without losing what the model actually needs to know.

The official repo for "LLoCo: Learning Long Contexts Offline"

Pytorch implementation for "Compressed Context Memory For Online Language Model Interaction" (ICLR'24)

Context compression plugin for Claude Code. Trims large JSON, logs, stack traces, and source files before they enter the context window.

State aware knowledge compression, ingestion, and hybrid retrieval engine. Zero dependencies. Sub-100ms queries.

Zed extension for Headroom — context compression for AI agents

Cursor uses AI to edit code — we use AI to edit AI's context. 🪆 Context map + compression + version control for LLM context windows.

⚡ Cut Claude token usage by 90%+ — free, open-source, local-first context compression for Claude Code. Hybrid RAG (BM25 + ONNX vectors), AST chunking, reranking. No API needed.

LLM reliability layer -keeps agents alive with smart routing, context compaction, and local fallback

Local-first Model Context Protocol (MCP) memory layer for Codex CLI/Desktop, Claude Code, Gemini CLI, Qwen/DeepSeek/Ollama and agent workflows. SQLite + FTS5 compact context packs, token savings, read-only mode, no external memory server.

Convert long AI conversations into portable conversation state graphs for LLM handoffs.

Unified agent memory and context compression stack for 2026 NVIDIA + edge (Vera CPU, Grace, Jetson Thor, 3090). Glues busyBee-cpu, honey-comb, and rust-brain. Better effective reasoning per token.

Local streaming reverse proxy between AI coding agents (Claude Code, Cursor, Codex) and model APIs (Anthropic, OpenAI, Gemini, MiniMax). Meters every token + USD cost, compacts bloated context to cut pay-per-token API spend, and runs shadow-eval to prove quality held. ccusage-style metering + live local dashboard.

🦞 LobsterPress（龙虾饼） - Cognitive Memory System for AI Agents 基于认知科学的 LLM 永久记忆引擎

Rolling context compression for Claude Code — never hit the context wall. Auto-compresses old messages while keeping recent context verbatim. Zero config, zero latency. Works as a Claude Code plugin.