mcp-compress
Provides capabilities for lossless compression and decompression of Markdown documents to save space while maintaining data integrity.
mcp-compress
The first MCP server for data compression. Gives any AI agent the ability to compress, decompress, analyze, and store data.
10,000+ MCP servers exist. Zero for compression. This is the first.
Zero dependencies. Pure Node.js. Lossless round-trip. Auto-picks the best algorithm.
Benchmarks
Real results on real data types:
Data Type | Original | Compressed | Ratio | Saved |
Markdown docs (15KB) | 31.2 KB | 0.5 KB | 60.7x | 98.4% |
Repeated config (2KB) | 5.4 KB | 0.1 KB | 51.9x | 98.1% |
SQL query results (8KB) | 18.9 KB | 0.6 KB | 30.4x | 96.7% |
Log files (20KB) | 33.3 KB | 1.7 KB | 19.9x | 95.0% |
JSON API response (10KB) | 26.7 KB | 2.6 KB | 10.2x | 90.2% |
Time-series prices (4KB) | 20.5 KB | 3.0 KB | 6.9x | 85.5% |
CSV data (5KB) | 8.1 KB | 2.4 KB | 3.4x | 70.5% |
Every compression is lossless — decompress returns the exact original, byte-for-byte.
Related MCP server: shorthand-mem
Install
Claude Code
Add to ~/.claude/settings.json:
{
"mcpServers": {
"compress": {
"command": "npx",
"args": ["-y", "mcp-compress"]
}
}
}OpenClaw / Any MCP Client
npx mcp-compressSpeaks MCP protocol over stdio. Works with any MCP-compatible AI agent.
From Source
git clone https://github.com/ShipItAndPray/mcp-compress.git
cd mcp-compress
node index.jsTools
7 tools available to any connected agent:
Tool | What it does |
| Compress text/JSON/CSV. Auto-picks best algorithm (gzip, brotli, deflate). Returns base64 + ratio. |
| Decompress back to original. Lossless round-trip verified. |
| Shannon entropy, compressibility rating, all algorithms compared, recommendation. |
| Compress and persist to disk with a key. Compressed key-value store for agents. |
| Decompress and return stored data by key. |
| List all stored items with sizes and compression ratios. |
| Total items stored, bytes saved, overall compression ratio. |
Usage Examples
Compress a large API response:
compress(data: "<10KB JSON>", algorithm: "auto")
→ { ratio: "10.2x", saved_percent: "90.2%", algorithm: "brotli" }Analyze before compressing:
analyze(data: "<your data>")
→ { compressibility: "HIGH", best_ratio: "30.4x", recommendation: "compress everything" }Store data for later retrieval:
store(data: "<research notes>", name: "market-analysis")
→ { key: "market-analysis", ratio: "8.3x", saved: "12,450 bytes" }
retrieve(key: "market-analysis")
→ { data: "<original research notes>" }Check what you've stored:
stats()
→ { stored_items: 14, total_saved_bytes: 284102, overall_ratio: "11.2x" }Why This Exists
AI agents generate and consume massive amounts of text — API responses, code, docs, data
Context windows are expensive. Compressed storage = more data in less space = lower cost.
MCP is the standard protocol for AI agent tools. 10,000+ servers, none for compression.
Auto-algorithm selection means the agent doesn't need to know anything about compression — it just works.
How It Works
Auto-algorithm selection — tests gzip, brotli, and deflate on your data, picks the smallest result
Brotli wins 90% of the time — purpose-built for text, consistently 20-40% smaller than gzip
Compressed key-value store —
store/retrievegives agents persistent compressed storage at~/.mcp-compress/Shannon entropy analysis —
analyzetells you if compression is even worth it before you do it
Test Results
10/10 evals passing:
✓ Initialize returns protocol version
✓ Lists all 7 tools
✓ Compress returns valid base64 and ratio > 1x
✓ Round-trip is lossless
✓ Analyze returns compressibility recommendation
✓ Store and retrieve preserves data
✓ Stats returns valid counts
✓ List shows stored items
✓ Auto picks smallest algorithm
✓ Handles 100KB+ dataLicense
MIT
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseBqualityDmaintenanceAdvanced Token-Optimized Object Notation MCP server that compresses JSON with up to 85% token reduction using AI-powered pattern detection, providing lossless compression and decompression through 12 MCP tools.129MIT
- Flicense-qualityCmaintenanceSemantic compression MCP server that reduces document token usage by 60-80% using structured symbolic notation, enabling Claude to efficiently store, retrieve, diff, and summarise documents across PRD, CODE, PAPER, and MEETING domains.
- FlicenseBqualityCmaintenanceLocal MCP server for token optimization, providing tools to compress code/JSON, optimize prompts, and manage placeholder-based content redaction and hydration to reduce LLM token usage.5
- AlicenseAqualityAmaintenanceProvides reversible context compression for AI agents, reducing token usage while preserving the ability to retrieve original content, and serves as an MCP server for integration with tools like GitHub Copilot and Claude Code.31Apache 2.0
Related MCP Connectors
Zero-auth MCP: image optimize, cited storage/format data, and dev utilities LLMs get wrong.
AI Reasoning Cache & Consensus Layer with 11 MCP tools via Streamable HTTP.
MCP server for URL shortening and management
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/ShipItAndPray/mcp-compress'
If you have feedback or need assistance with the MCP directory API, please join our Discord server