📡 MCP Radar

AlexanderBoger/SaveContext

Git LFS for your LLM context — 20x Context Savings — an MCP server that keeps big documents out of the model and returns byte-exact quotes for a fraction of the tokens.

5 stars
0 forks
Python
momentum ▲ 10.0
created 2026-07-23
on radar since 2026-07-25
star trend 4 → 5 since 2026-07-25
anthropicclaudecontext-compressionllmmcpmcp-servermodel-context-protocolpythonragtoken-optimization
View on GitHub ↗Homepage ↗

About SaveContext

Keep big documents outside the model — hand it a small searchable handle instead. The model reads a cheap summary to find what it needs, then pulls back the exact original words to quote, using a fraction of the tokens.

An MCP server for Claude Code, Claude Desktop, and any Model Context Protocol client — with a standalone CLI too.

With uv, add SaveContext to Claude Code in one line — no clone, no pip, no virtualenv:

Then open Claude Code, run /mcp to confirm the tools loaded, and point the model at a big document and just ask — it ingests, looks up, and quotes on its own. To pin where the database lives, add -e SAVECONTEXTDB=/abs/path.db before the --.

Using a different MCP client (Claude Desktop, etc.)? Point its server config at the same command: uvx savecontext. Prefer no MCP at all? The CLI does everything standalone.

Most context tools are lossy summarizers — helpful, but they blur the exact numbers and can't tell you when they're guessing. SaveContext does the two things they can't:

- 🎯 Exact, with receipts. Facts come back as the original characters with their source location, never a paraphrase — every span reconstructs byte-for-byte. A summary is for finding; the quote is for trusting. - 🙅 Says "not found" instead of inventing. When the answer isn't in the document, retrieval is flagged weak and the agent abstains — rather than hallucinating a plausible one. - 📉 Flat cost. The payload to answer questions stays ~7,300 tokens whether the document is 15k or 150k. Pasting scales linearly; this doesn't. - 🔁 One call, many questions. lookup answers a batch of quer

From the project README.

Maintaining this server?

Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:

[![On MCP Radar](https://mcp.liqiwa.com/badge.svg)](https://mcp.liqiwa.com/s/AlexanderBoger--SaveContext.html)

Related MCP servers

chris-asmussen/warden

One MCP server that fronts many MCP servers + Skills behind a small tool set and routes to the best one, keeping an agent's context tiny. Migrate your Claude Code MCPs/skills behind it.

⭐ 10Python
KernelLord/pickysteve

Skill router and context picker for Claude Code, Cursor, and Aider. Auto-picks the right skill for every prompt, with prompt-injection filtering for MCP.

⭐ 6Python
vinodborole/okf-kit

Turn any website into a portable, agent-ready Open Knowledge Format (OKF) bundle — no LLM required. Crawl, sync, chat offline, and serve docs to coding agents over MCP.

⭐ 5Python
xDarkzx/SongForge-MCP

Generate original AI music tracks via Claude Desktop — MCP server powered by ACE-Step 1.5, with vocal/instrumental stem splitting and reference-style matching. No music software or AI/ML knowledge nee

⭐ 4Python
AdamRussak/self-doc

Self-hosted documentation RAG pipeline that serves semantic search to LLM agents over the Model Context Protocol.

⭐ 3Python
TeodorMCP/universal-connector-mcp

Universal MCP server: connect any OpenAPI/Swagger, GraphQL, gRPC or SOAP API to AI agents. Security-first, local, token-efficient. Any API. One server.

⭐ 3Python

📬 Get the weekly radar in your inbox

The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.