context-compression 开源项目推荐（396 个）

open-compress /claw-compactor

14-用于LLM令牌压缩的Fusion Pipeline阶段-可逆压缩、AST感知代码分析、智能内容路由。零LLM推理成本。麻省理工学院获得许可。

ojuschugh1 /sqz

压缩LLM上下文以节省代币并降低成本

juyterman1000 /entroly

在AI编码上，将你的Claude / OpenAI / Gemini账单砍70%到95%。本地代理，压缩上下文，保持提供者缓存热，并验证LLM输出（免费幻觉保护）。Cursor、Claude Code、Codex、Aider + 34 个额外和自定义服务商的直接访问 — 30 秒，代码不变

activeing123 /mcptoon

One zero-dependency MCP gateway CLI for every server and agent skill. Token optimization, tool discovery and context compression: 71,929 -> 581 tokens (-99.2%, measured), schemas stay out of context. One config for Claude Code, Codex, Cursor, every agent. 561KB, pure Python. | 零依赖 CLI：管所有 MCP 工具与技能，工具发现省 99.2% token。

fajarhide /contexere

The one place for the skills, keys and memory of every agent you use. Formerly OMNI.

HoangP8 /tokless

一个统一的工具包，用于高效且有效的编码代理（Karpathy原则、Caveman、Ponytail、TEK、CodeCurve、Context-Mode）。最短设置时间不超过30秒。任何操作系统。

snap-stanford /memorilla

🦍 Memorilla: Latent Semantic Memory for LLMs

ZipengWu365 /TOMC

TOMC Memory: CPU context compression and task memory for Codex, Claude Code, Cursor, TRAE, WorkBuddy and other MCP clients.

Ahren09 /SARA

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

LearnPrompt /cc-harness-skills

受CC启发的便携技能，用于内存、验证、多代理协调、上下文压缩和主动编码代理工作流程。

255308153 /CtxGuard

面向 AI Agent 的高并发、低延迟上下文治理与 Prompt Cache 守护网关 (Tree-sitter AST / Tool Delta / 50%~80% Token 削减)

lennney /mcp-slim-guard

Context compression for MCP. Same upstream call, exact results recoverable.

borhen68 /TokenTamer

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.

jeffreysijuntan /lloco

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

chopratejas /headroom-zed

Zed extension for Headroom — context compression for AI agents

PCIRCLE-AI /toonify-mcp

Context compression plugin for Claude Code. Automatically trims large tool output—JSON, YAML, stack traces, and logs—before it enters the context window.

khanhtq29vn /contextcraft

AST-Aware, Personalized-PageRank-Ranked, Token-Budgeted Codebase Compressor for LLMs

snu-mllab /Context-Memory

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

castnettech /mnemosyne

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

susyimes /LAIN-agreement

铃音的记忆写入协议 - an informatics protocol: fragmented, cue-based, symbolic, highly compressed information for AI agents; usable for memory writing and transmission, decodable approximately or via retrieval.