Context Window Python Packages

headroom-ai

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.

jcodemunch-mcp

Cut AI token costs 95%+ on code exploration. The leading MCP server for precise, symbol-level GitHub code retrieval via tree-sitter AST. Works with Claude Code, Cursor & any MCP client. 313B+ tokens saved.

cocoindex-code

A super light-weight embedded code search engine CLI (AST based) that just works - improves speed and efficiency for coding agent 🌟 Star if you like it!

diffctx

Smart git diff context for LLMs and AI agents: budgeted fragment selection, deterministic, MCP server included

entroly-core

Cut AI context cost without trusting the compressor. Every reduction is reversible, byte-exact recoverable, and carries an auditable receipt. Local-first, works through proxy, MCP, SDK, or agent wrapper.

kijito-tools

Run Claude Code sessions that know their own context and can run themselves: real context self-check, armed-pane auto-catch-up, and a safe self-clear loop. Kijito-backed, works standalone.

tokenpak

Local proxy that puts the right project context in front of the model before it starts working — with a receipt for what was added

ctxfw

High-assurance in-memory Tree-Sitter AST context firewall and pruning MCP server for coding agents (-72.4% token mass).

entroly

Cut AI context cost without trusting the compressor. Every reduction is reversible, byte-exact recoverable, and carries an auditable receipt. Local-first, works through proxy, MCP, SDK, or agent wrapper.

contextrot

Find out where your coding agent starts degrading. Personal context-rot analytics from your own sessions - 100% local, zero-config: uvx contextrot

token-save-mcp

Stop your coding agent burning context on file reads. Delegates big reads to a cheap worker model and returns only the answer — 87-99% saved, measured per call. A hook enforces it in Claude Code; the MCP tools work in any client.

contextzip

CLI tool for intelligent project packaging with framework detection, smart exclusions, and git-aware file selection.

kopipasta

`cat project | LLM | patch`. Transparent context control and interactive patching for terminal-centric LLM workflows.

tzero-mcp

AST Codebase Context Engine & 21-Tool MCP Server for AI Coding Agents — Up to 95% Token Reduction & Zero-Leak Security Architecture

sqz

Context compression for AI coding agents: compresses tool output before it enters the model, dedups repeats to 13-token refs. Claude Code, Cursor, Codex, Kiro, Zed, any MCP client. Rust, zero LLM calls.

latent-gate

MCP server that shrinks logs, JSON, docs and web pages before your AI reads them. Offline, deterministic, ~3ms, no API key. 66% fewer tokens on a mixed dev corpus, 90%+ on logs.

context-guardian

Compaction for local models that actually fires and never bricks the session. OpenAI-compatible proxy + DeepSeek Harness engine.

memshelf-mcp

Working memory for AI agents: digest-indexed episodes on a git shelf, recalled a section at a time.

gaffr

Context-lifecycle handler for coding agents: counters, cadenced re-injection, prime sections

tool-guardian

Seven MCP servers for 934 tokens instead of 20,005 (measured). MCP router + native DeepSeek Harness bundle.

contextburn

How much of what your coding agent spends is actually work: the share of paid tokens that became output, not context re-reading. Local, no network calls.

a2cr-mcp

MCP server for client-encrypted AI agent handoffs with WorkBaton and WorkStash

lockkeeper

Skill router and prompt-injection firewall for AI coding agents: gives Claude Code, Codex and Cursor the few skills, MCP servers and tools each task needs.

hilbras-remembra

🧠 External memory for AI assistants — facts, decisions, roles and history stored outside the context window. MCP server for OpenCode, Claude Code, Cline, Kimi Code.