context-pruning 开源项目推荐（22 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。

Sev7eEn7 /dsh-sieve

dsh-sieve: context engineering & token optimization plugin for DeepSeek Harness (DSH) — tool output filtering, context pruning, progressive skill disclosure. 36% smaller payload in offline replay. DSH 上下文管理与 token 优化节省插件。

AshishKumar4 /better-compact

Staged context pruning for OpenCode, Oh My Pi, pi, and Claude Code.

umitkacar /llm-context-optimizer

Biological code organization system with 1,029+ production-ready snippets - 95% token reduction for Claude/GPT with AI-powered discovery & offline packs

pi-vault /pi-dcp

Pi extension for dynamic context pruning

JimmyWesley /rlcd-gateway

Self-hosted gateway for LLMs and decision models. Claude Code, Codex, OpenCode and any OpenAI/Anthropic SDK app reach any provider (OpenRouter, Groq, Ollama…) with context pruning; Jev and open-rlcd System One decisions get audit and calibration. Live dashboard, single Go binary.

AZagatti /trimwire

Local gateway that prunes Claude Code's API context in-flight — dedup, purge failed calls, trim bloat, strip old screenshots. Anthropic-sanctioned (ANTHROPIC_BASE_URL), no CA cert, no restart.

8b-is /longrun-eval-kompress

the research synthethizer outer loop

kevinpita /pi-jev-context

Reversible context pruning for Pi, powered by TypeSafe Jev. Keep useful context without deleting session history.

ivanalvarizjr /CodeBOT

You approve the plan. CodeBOT does the rest. Semi-autonomous multi-agent development: plan first, execute fully, report every change. Currently built for Copilot & Codex.

codesbykhairannoor /capybala

High-Speed, Zero-Dependency Cognitive Deadlock & Oscillation Interceptor for Autonomous AI Coding Agents.

CompleteTech-LLC-AI-Research /jev-context-fabric

Local-first, source-backed memory layer for coding agents with an installer for eight harnesses and reversible /prune

chiro2001 /dsh-dcp

DeepSeek Harness 动态上下文管理插件（Dynamic Context Pruning for dsh），对标 opencode-dcp

CompleteTech-LLC-AI-Research /jev-prune-kit

Capability-aware context-pruning installer and Jev assessment engine for coding agents; experimental, not live-host tested

jaydeeep5764 /antiecho

Stop LLM context echoing, repetition spirals, and memory poisoning in multi-turn chats.

Helix5240 /The-Education-Tutor-For-Remote-India

AI-powered tutoring system for Indian state-board students. Upload textbook PDFs and get curriculum-aligned answers instantly. Uses Context Pruning to score and filter chapters before querying Gemini LLM — reducing API costs by ~80%. Built with Python, Flask, FAISS, sentence-transformers, and Gemini 2.5 Flash.

ZeeshanGeoPk /ollama-rag-memory

Local Ollama RAG memory proxy using ChromaDB and nomic-embed-text to retrieve relevant chat history, compress context, and reduce LLM tokens.

theroich /opencode-jev-pruner

OpenCode V2 plugin: dynamic context pruning for MCP tools and skills using Jev (TypeSafe System One)

bawfng04 /ContextSieve

Enterprise-grade Context Pruning for RAG. Hierarchical 2-stage architecture (MiniLM-L6 ONNX INT8) achieving 60-75% token compression at <20ms latency.

augustoolucas /fast-opencode-compaction

opencode V2 adapter for fast-jev-compaction: prunes stale tool calls and results from the outgoing request using a typed-decision endpoint