mcp-crush
Integrates with Ollama by using OLLAMA_HOST, probing the model context window, and tuning truncation/schema packing to keep MCP tool catalogs and responses within smaller context budgets.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcp-crushaudit my MCP config and show token usage per server"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
mcp-crush
One line: npx mcp-crush wrap -- <your-mcp-cmd> (or bunx mcp-crush wrap -- <your-mcp-cmd>): stdio proxy that shrinks MCP tool catalogs and prunes oversized JSON responses by 60-85%.
Why
Large MCP servers (GitHub, Cloudflare, Notion) dump thousands of tokens into tools/list before your agent does any work. Local 8k models cannot fit them, and cloud APIs bill you for that handshake on every session.
mcp-crush sits in the middle using the same JSON-RPC wire protocol while emitting smaller payloads.
Related MCP server: mcp-lazy-proxy
Install
bunx mcp-crush audit
npx mcp-crush audit
# or install in your project / globally
bun add -d mcp-crush
npm i -D mcp-crushJSR:
bunx jsr add @kanchingyg/mcp-crush
npx jsr add @kanchingyg/mcp-crushQuick start
# Inspect token overhead from your MCP config
npx mcp-crush audit
bunx mcp-crush audit
# Wrap any stdio MCP server
npx mcp-crush wrap -- npx -y @modelcontextprotocol/server-github
# Ollama-friendly schema pack (~600 token budget)
npx mcp-crush wrap --ollama-pack -- npx -y @modelcontextprotocol/server-github
# Expose audit/compress/prune as MCP tools
npx mcp-crush --mcpOllama / LM Studio
Set OLLAMA_HOST (default http://127.0.0.1:11434). With --mcp, mcp-crush probes n_ctx and tunes truncation automatically. Use --ollama-pack on wrap to cap tools/list schemas around 600 tokens.
MCP Tools Reference
Tool | Mutates | Description |
|
| Parse an MCP configuration file and measure per-server |
|
| Minify an MCP |
|
| Strip redundant UI metadata ( |
|
| Persist a compressed MCP tool profile to disk (requires |
Editor and Agent Integration
Cursor (.cursor/mcp.json)
{
"mcpServers": {
"mcp-crush": {
"command": "npx",
"args": ["-y", "mcp-crush", "--mcp"]
}
}
}Claude Code CLI
claude mcp add mcp-crush -- npx -y mcp-crush --mcpProgrammatic Usage (JSR / TypeScript)
import * as bridge from "jsr:@kanchingyg/mcp-crush/bridge";Runtime Safety and Guardrails
Stdio frame isolation:
--mcpredirects strayconsole.logtostderrso JSON-RPC onstdoutstays valid.Read-only default: Mutating tools need
--allow-write;--dry-runblocks writes.Loop breaker: Identical tool calls halt after three repeats (SHA-256 of name + args).
Context budget:
wrap/ MCP proben_ctxand head/tail-truncate large tool output.Config path sandbox:
audit_mcp_configresolvesconfigPathunder the process cwd.Secret redaction: AWS keys, bearer tokens, and
sk-*patterns scrubbed from tool text.
License
This server cannot be deployed
Maintenance
Related MCP Connectors
Security & DLP proxy for MCP: tool-poisoning scans, PII redaction on tool args/results. Beta.
Remote MCP for MCP tool deprecation receipt, structured receipts, audit logs, and reviewer-ready evi
Governed MCP gateway: one endpoint for your tools, with credential custody and audit log.
Statically audits MCP tool surfaces for token cost, schema quality, and design issues.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables analyzing and optimizing context window token usage by measuring, compressing, and eliminating waste from MCP tool definitions.27 npmMIT
- AlicenseNot gradedqualityFmaintenanceReduces MCP tool schema token overhead by lazy-loading and caching schemas, with response compression and verifiable savings metrics.15 npm1MIT
- AlicenseNot gradedqualityCmaintenanceEnables enterprise MCP security auditing through 12 deterministic no-LLM tools for evidence-gated claims, skill/prompt supply-chain audits, token profiling, and server auth-mode checks.MIT
- AlicenseNot gradedqualityBmaintenanceEnables unified access to multiple MCP backend servers through one local daemon, with curated per-client tool profiles, allow/deny policies, audit logging, and drift protection.13 npm1MIT