QWED-MCP
Enables deterministic verification for AI assistants by executing Python code that uses symbolic engines like SymPy and Z3 for math, logic, and code analysis.
README
<div align="center"> <img src="assets/logo.svg" alt="QWED Logo - AI Verification Engine" width="80" height="80"> <h1>QWED-MCP š</h1> <h3>Deterministic Verification for Claude Desktop & VS Code</h3>
</div>
MCP Server for QWED Verification ā Bring deterministic verification to Claude Desktop, VS Code, and any MCP-compatible AI assistant.
š Full Documentation: docs.qwedai.com/mcp
ā” Quick Install
pip install qwed-mcp
š Setup with Claude Desktop
Step 1: Find your config file
| OS | Path |
|---|---|
| Windows | %APPDATA%\Claude\claude_desktop_config.json |
| macOS | ~/Library/Application Support/Claude/claude_desktop_config.json |
| Linux | ~/.config/Claude/claude_desktop_config.json |
Step 2: Add QWED-MCP
macOS/Linux:
{
"mcpServers": {
"qwed-verification": {
"command": "qwed-mcp"
}
}
}
Windows (use python -m):
{
"mcpServers": {
"qwed-verification": {
"command": "python",
"args": ["-m", "qwed_mcp.server"]
}
}
}
Step 3: Restart Claude Desktop
Quit completely (system tray ā Quit) and reopen.
Step 4: Test it!
Ask Claude:
"Write a python script that verifies a 10,000 investment at 7.5% for 5 years using the
qwed_newmath engine, and run it usingexecute_python_code."
ā ļø Migration Note: Deprecation of verify_* Tools
To solve "context bloat" and align with the new MCP standard (RFC-9728), all 1:1 functional tools (e.g., verify_math, verify_sql, verify_code) have been removed as of v0.2.0.
They have been replaced with a single, highly capable tool:
š execute_python_code
Before:
"Use
verify_mathto check this formula." (Claude loads 14 different tool schemas into context)
After:
"Use
execute_python_codeto write and run a script that importsqwed_new.engines.math_engineto verify..." (Claude loads 1 tool schema into context)
If you see an "Unknown tool" error, it means Claude is trying to use a legacy tool. Simply tell Claude: "The verify_* tools are removed. Use execute_python_code to natively write and run a Python verification script."
š§ Available Tools
| Tool | Description | Use Case |
|---|---|---|
execute_python_code |
Subprocess Execution | The single entrypoint for all QWED capabilities. Executes dynamically generated Python code in a subprocess with restricted environment variables. Note: Runs with server privileges; ensure inputs are trusted. |
š” Example Prompts for Claude
Note: Claude already knows how to use QWED natively via standard Python imports.
Financial Calculations
A bank says: "Invest $10,000 at 7.5% compounded quarterly for 5 years = $14,356.29"
Please write a short Python script using the standard compound interest formula to verify this, and run it with execute_python_code.
Loan EMI Verification
Verify: ā¹10,00,000 loan at 9% for 5 years = EMI of ā¹20,758
Write a python script importing necessary tools to verify this EMI calculation, and execute it using execute_python_code.
Complex Reasoning Workflows (The Power of Python)
Read the user terms in the attached document.
1. Use execute_python_code to extract and verify the legal clauses using qwed_legal.
2. In the same script, verify if the referenced financial penalties align with the allowed boundaries.
šļø How It Works
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
ā Claude Desktop / VS Code ā
ā (MCP Client) ā
āāāāāāāāāāāāāāāāāāā¬āāāāāāāāāāāāāāāāāāāāāāāāāā
ā MCP Protocol (JSON-RPC)
ā¼
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
ā QWED-MCP Server ā
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā¤
ā execute_python_code() ā
ā āāāŗ Subprocess Execution (Restricted Env)ā
ā āāāŗ Native QWED library execution ā
āāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāāā
šÆ Why QWED-MCP?
Note: Subprocess execution provides answers/checks purely based on what QWED SDK methods are invoked inside the executed scripts. Execution itself does not guarantee injection detection without specific SDK calls.
| Without QWED-MCP | With QWED-MCP |
|---|---|
| LLM calculates ā 95% correct | Executes Python script calling qwed_finance ā 100% correct |
| LLM writes SQL ā might inject | Script uses qwed_new analyzer ā injection detected |
| LLM reasons ā might be wrong | Z3 solver executed via SDK ā formally proven |
| LLM codes ā might be unsafe | AST check script executed ā security checked |
š” What QWED-MCP Is (and Isn't)
ā QWED-MCP IS:
- MCP Server that adds verification tools to Claude Desktop and VS Code
- Deterministic ā uses SymPy (math), Z3 (logic), AST (code) for exact verification
- Open source ā works with any MCP-compatible AI assistant
- A safety layer ā catches LLM hallucinations in real-time
ā QWED-MCP is NOT:
A replacement for Claudeā it enhances Claude with verification toolsA chatbotā it's a backend server that Claude callsInternet-connectedā all verification happens locallyA fine-tuned modelā uses symbolic engines, not ML
Think of QWED-MCP as giving Claude a "calculator" for math and a "theorem prover" for logic.
Claude reasons. QWED-MCP verifies.
š How We're Different from Other MCP Servers
| Aspect | Other MCP Servers | QWED-MCP |
|---|---|---|
| Purpose | Connect to APIs, databases, files | Verify LLM outputs |
| Approach | Fetch external data | Compute deterministic proofs |
| Engines | API wrappers | SymPy, Z3, AST analyzers |
| Accuracy | Depends on data source | 100% mathematically proven |
| Offline | Often need internet | Fully local, no APIs |
With Claude Desktop
āāāāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāāāā
ā Claude ā ā QWED-MCP ā ā Verified Answer ā
ā "What's d/dx x³?" ā āāāŗ ā execute_python_code() ā āāāŗ ā "3x²" ā ā
ā "Write script to ā ā Runs SymPy natively ā ā (STDOUT Captured) ā
ā check." ā āāāāāāāāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāāāā
āāāāāāāāāāāāāāāāāāāāā
š Security & Privacy
All verification happens locally. Nothing is sent to external servers.
| Concern | QWED-MCP Approach |
|---|---|
| Data Transmission | ā No external API calls |
| Storage | ā Nothing logged or stored |
| Dependencies | ā Local engines (SymPy, Z3) |
| Code Analysis | ā Your code never leaves your machine |
Perfect for:
- Enterprises with strict security policies
- Air-gapped development environments
- Sensitive code review workflows
ā FAQ
<details> <summary><b>Is QWED-MCP free?</b></summary>
Yes! Open source under Apache 2.0. Use it commercially, modify it, distribute it. </details>
<details> <summary><b>Does it work with VS Code Copilot?</b></summary>
QWED-MCP works with any MCP-compatible client. VS Code with Claude extension supports MCP, so yes! </details>
<details> <summary><b>Do I need an API key?</b></summary>
No. QWED-MCP runs entirely locally. No API keys, no cloud calls. </details>
<details> <summary><b>What's the difference between this and QWED-Core?</b></summary>
QWED-Core is the Python library. QWED-MCP wraps it as an MCP server so Claude can use it as a tool. </details>
<details> <summary><b>Can I add my own verification tools?</b></summary>
Yes! The server is extensible. Fork it and add your custom @mcp.tool() functions.
</details>
šŗļø Roadmap
ā Released (v0.2.0)
- [x] Context bloat resolution (RFC-9728 compatibility)
- [x] Unified
execute_python_codeenvironment - [x] Secure process isolation (env-restricted) and robust timeouts
- [x] Claude Desktop integration
- [x] Windows/macOS/Linux support
š§ In Progress
- [ ]
verify_jsonā JSON Schema validation tool - [ ]
verify_financeā NPV/IRR/amortization tool - [ ] Cursor IDE integration guide
š® Planned
- [ ]
verify_legalā Deadline and liability verification - [ ]
verify_statisticsā Hypothesis test validation - [ ] SSE (Server-Sent Events) transport for web UIs
- [ ] TypeScript implementation
š Examples
See the examples/ folder for:
- Python client usage
- Sample verification scripts
- Integration examples
š ļø Development
# Clone
git clone https://github.com/QWED-AI/qwed-mcp.git
cd qwed-mcp
# Install with dev dependencies
pip install -e ".[dev]"
# Run tests
pytest tests/ -v
# Format code
black src/
š Documentation
| Resource | Link |
|---|---|
| Full Docs | docs.qwedai.com/mcp |
| Tools Reference | docs.qwedai.com/mcp/tools |
| Examples | docs.qwedai.com/mcp/examples |
| Troubleshooting | docs.qwedai.com/mcp/troubleshooting |
| MCP Protocol | modelcontextprotocol.io |
š Related Projects
- QWED Core ā github.com/QWED-AI/qwed-verification
- QWED-UCP ā github.com/QWED-AI/qwed-ucp
- QWED Open Responses ā github.com/QWED-AI/qwed-open-responses
š License
Apache 2.0 ā See LICENSE
<p align="center"> <b>Built by <a href="https://qwedai.com">QWED AI</a></b><br> <i>Making AI outputs trustworthy through formal verification</i> </p>
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.