Content
# thoughtproof-mcp
[](https://www.npmjs.com/package/thoughtproof-mcp)
[](https://github.com/ThoughtProof/thoughtproof-mcp/actions/workflows/ci.yml)
[](https://opensource.org/licenses/MIT)
MCP server for [ThoughtProof](https://thoughtproof.ai) — verify AI reasoning with adversarial multi-model consensus.
3–4 LLMs (Grok, Gemini, DeepSeek, Sonnet) independently evaluate every claim. A dedicated red-team model critiques their verdicts. A synthesizer (Sonnet) weighs everything and returns **ALLOW**, **BLOCK**, or **UNCERTAIN** with confidence score and objections.
## Quick Start
```json
{
"mcpServers": {
"thoughtproof": {
"command": "npx",
"args": ["-y", "thoughtproof-mcp"],
"env": {
"THOUGHTPROOF_API_KEY": "tp_op_your_key_here"
}
}
}
}
```
Works with **Claude Desktop**, **Cursor**, **Windsurf**, **Cline**, and any MCP-compatible client.
## Tools
### `verify_claim`
Verify any claim or AI-generated reasoning before acting on it.
| Parameter | Type | Default | Description |
|-----------|------|---------|-------------|
| `claim` | string | *(required)* | The text to verify |
| `stakeLevel` | `low` / `medium` / `high` / `critical` | `medium` | Risk level — higher stakes trigger deeper verification |
| `domain` | `financial` / `medical` / `legal` / `code` / `general` | `general` | Domain context for specialized verification |
| `speed` | `fast` / `standard` / `deep` | `standard` | Verification depth |
### `check_agent_score`
Look up an agent's composite trust score on the ERC-8004 registry.
| Parameter | Type | Description |
|-----------|------|-------------|
| `agentId` | string | Agent ID to look up |
| `domain` | string | Optional domain filter |
## Example
In Claude Desktop or Cursor, just ask:
> "Verify the claim: GPT-5 achieves 95% accuracy on MMLU-Pro"
The tool returns:
```
⚠️ UNCERTAIN (42% confidence)
Claim: "GPT-5 achieves 95% accuracy on MMLU-Pro"
Objections:
- Insufficient public benchmark data to confirm
- Historical accuracy claims have been overstated
- MMLU-Pro methodology has known ceiling effects
⚡ 3.2s | Adversarial Multi-Model Consensus
```
## How It Works
```
Your AI Agent
│
▼
┌──────────────────┐
│ thoughtproof-mcp │ ← MCP Server (this package)
└──────────────────┘
│
▼
┌──────────────────┐
│ ThoughtProof API │ ← api.thoughtproof.ai (RV)
└──────────────────┘
│
▼
┌───────────────────────────────────────────┐
│ Stage 1: Independent Evaluation │
│ 3–4 LLMs (Grok, Gemini, DeepSeek, │
│ Sonnet) each examine the claim │
│ │
│ Stage 2: Red-Team Critique │
│ 1 dedicated model challenges all │
│ initial verdicts │
│ │
│ Stage 3: Synthesis │
│ Sonnet weighs verdicts + critique │
│ → final decision │
└───────────────────────────────────────────┘
│
▼
ALLOW / BLOCK / UNCERTAIN
+ confidence % + objections
```
## Pricing
| Speed | Models | Cost per verification |
|-------|--------|-----------------------|
| fast | 2 | $0.008 |
| standard | 4 | $0.02 |
| deep | 5+ | $0.08 |
Payment: API key (operator account) or x402 micropayment (USDC on Base).
## API Key
Get an operator API key at [thoughtproof.ai](https://thoughtproof.ai). Without a key, verifications use x402 micropayments automatically.
## Configuration
| Environment Variable | Default | Description |
|---------------------|---------|-------------|
| `THOUGHTPROOF_API_KEY` | *(none)* | Operator API key |
| `THOUGHTPROOF_BASE_URL` | `https://api.thoughtproof.ai` | API base URL |
## Development
```bash
git clone https://github.com/ThoughtProof/thoughtproof-mcp.git
cd thoughtproof-mcp
npm install
npm run build
npm test
npm run dev # Run with tsx (hot reload)
npm run inspect # Test with MCP Inspector
```
## Related
- [ThoughtProof](https://thoughtproof.ai) — Decision verification for AI agents
- [pot-cli](https://github.com/ThoughtProof/pot-cli) — CLI for reasoning verification
- [ERC-8004](https://eips.ethereum.org/EIPS/eip-8004) — Autonomous Agent Registry
## License
MIT — [ThoughtProof](https://thoughtproof.ai)
