Scorechain MCP

Connect any MCP-compatible AI client to Scorechain's blockchain compliance intelligence. Score addresses, trace fund flows, investigate transactions, and search your Scorechain data — all from your assistant, over a single hosted endpoint.

https://mcp.scorechain.com

AI client connecting to the Scorechain MCP endpoint, which calls the Scorechain API

Authentication

API Key

Send your Scorechain API key in the X-API-KEY header on every request. Get your key from the Scorechain dashboard.

POST https://mcp.scorechain.com
X-API-KEY: ...
Content-Type: application/json

Treat the key like a password. Store it in your client's secret store or an environment variable — never in source control.

Clients that support MCP OAuth let you connect by signing in to Scorechain in your browser — no API key to copy or store. The client discovers the authorization server from:

https://mcp.scorechain.com/.well-known/oauth-protected-resource

It opens a browser, you sign in to Scorechain and approve access, and the client receives a short-lived token scoped to your account. See Use Scorechain in ChatGPT for a step-by-step example.

Connect Your Client

Configuration shape varies by host. The server URL and header stay the same everywhere.

Claude Desktop

Easiest: add it as a connector and sign in with your Scorechain account — see the dedicated step-by-step guide: Use Scorechain in Claude Desktop.

To configure manually with an API key instead, open Settings → Developer → Edit Config and add:

{
  "mcpServers": {
    "scorechain": {
      "type": "http",
      "url": "https://mcp.scorechain.com/mcp",
      "headers": {
        "X-API-KEY": "..."
      }
    }
  }
}

Restart Claude Desktop. Tools appear in the MCP tools list.

Claude Code

claude mcp add --transport http scorechain https://mcp.scorechain.com/mcp \
  --header "X-API-KEY: ..."

Cursor

Cursor connects over OAuth. Open Settings → MCP → Add new MCP server and add only the URL:

{
  "mcpServers": {
    "scorechain": {
      "url": "https://mcp.scorechain.com/mcp"
    }
  }
}

The first time you use the server, Cursor opens your browser to sign in to Scorechain and approve access. Once you complete the OAuth flow, the tools become available.

Cursor MCP settings panel with Scorechain configured

VS Code (GitHub Copilot Chat)

Add to .vscode/mcp.json in your workspace or user settings:

{
  "inputs": [
    {
      "type": "promptString",
      "id": "scorechain_key",
      "description": "Scorechain API key",
      "password": true
    }
  ],
  "servers": {
    "scorechain": {
      "type": "http",
      "url": "https://mcp.scorechain.com",
      "headers": {
        "X-API-KEY": "${input:scorechain_key}"
      }
    }
  }
}

ChatGPT

See the dedicated step-by-step guide: Use Scorechain in ChatGPT.

OpenCode

Add the following to your OpenCode configuration file (.opencode/config.json or equivalent):

{
  "mcp": {
    "scorechain": {
      "enabled": true,
      "type": "http",
      "url": "https://mcp.scorechain.com",
      "headers": {
        "X-API-KEY": "..."
      }
    }
  }
}

Other Clients

Any MCP client supporting remote HTTP transport works. Configure with:

Field	Value
URL	`https://mcp.scorechain.com/mcp`
Transport	HTTP (streamable)
Auth header	`X-API-KEY: ...`

Tools

All tools are read-only. They query your Scorechain account data in real time.

Category	Tool	Description
Core	`list_blockchains`	Supported blockchains and coverage levels
Core	`get_address`	Address metadata, entity labels, custom scores, cases, list memberships
Core	`get_transaction`	Transaction details by hash
Core	`score_address`	Real-time AML risk scoring
Visualization	`visualize_fund_flow`	Mermaid sankey diagram of fund flows
Visualization	`visualize_interactions`	Multi-hop interaction graph toward target entities or types
Intelligence	`get_address_transactions`	Paginated transaction history
Intelligence	`get_address_activity`	Balance and volume over time
Intelligence	`get_address_funder`	Original funder identification (provenance)
Intelligence	`get_asset_holders`	ERC-20 holder concentration and risk distribution
Search	`search`	Search addresses, transactions, entities, cases, labels, lists

Resources

Reference material the AI reads on demand to interpret results and select tools.

Resource URI	Purpose
`scorechain://capabilities`	Tool selection guide — which tool answers which question
`scorechain://entity-types`	Entity type reference for filters, score breakdowns, and interaction targets
`scorechain://risk-guide`	Risk score interpretation, severity levels, and decision framework
`scorechain://investigation-playbook`	Forensic playbook — laundering typologies, investigation workflows, tool sequences

Prompts

Workflow templates the AI can invoke as structured investigations. Each runs multiple tools in sequence and produces a formatted report.

Prompt	Purpose	Key inputs
`investigate-address`	Full forensic investigation — scoring, provenance, fund flows, network topology, behavioral analysis	`address`, `blockchain`
`trace-stolen-funds`	Trace stolen funds from source, map the laundering path, identify cash-out points for recovery	`address`, `blockchain`, `context` (optional)
`screen-wallet`	Fast accept/reject screening for onboarding, deposits, withdrawals, or counterparty checks	`address`, `blockchain`, `purpose` (optional)
`exposure-analysis`	Quantify direct and indirect exposure to high-risk entity categories	`address`, `blockchain`, `categories` (optional)
`cluster-investigation`	Determine if multiple addresses form a coordinated cluster controlled by one entity	`addresses` (comma-separated), `blockchain`

All prompts use only the 11 tools listed above. Output includes Mermaid diagrams (verbatim from tools) and structured reports.

How to Read Scores

Scorechain uses an inverse scale. Lower score = higher risk.

Score	Severity	Meaning
1	`CRITICAL_RISK`	Direct exposure to sanctioned or illicit entity
2–29	`HIGH_RISK`	Strong high-risk exposure
30–69	`MEDIUM_RISK`	Moderate exposure, enhanced due diligence warranted
70–100	`LOW_RISK`	Minimal exposure

A score of 1 is the worst possible result. Make sure your prompts and downstream logic account for this.

Example Prompts

Once connected, try these in your assistant:

"Investigate this Ethereum address — full forensic analysis with fund flows and interaction graphs."
"Screen this Bitcoin address for onboarding — should we accept or reject?"
"Trace where the stolen funds from this TRON address ended up. The exploit happened on 2024-03-15."
"Analyze this address's exposure to mixers, darknet, and sanctioned entities."
"Check if these 4 Ethereum addresses are part of the same cluster."
"Search Scorechain for this entity name and pick the most relevant match."

AI assistant running investigate-address and producing a structured risk report with Mermaid fund flow diagram

Connection Details

Property	Value
URL	`https://mcp.scorechain.com/mcp`
Transport	HTTP (streamable)
Auth	`X-API-KEY` header, or OAuth (sign in with Scorechain)

Troubleshooting

Tools don't appear in the client

Restart the client after editing config. Confirm the URL is exactly https://mcp.scorechain.com/mcp and that your client supports remote HTTP MCP transport.

401 Unauthorized

Your API key is wrong, expired, or not being sent. Verify the header name is X-API-KEY (spelling matters, case doesn't). Regenerate the key from the Scorechain dashboard if needed.

Connection timeout on long operations

Some clients enforce a tool timeout (default 60s in certain environments). If scoring or visualization calls time out, increase your client's MCP tool timeout setting. For example in Codex:

[mcp_servers.scorechain]
url = "https://mcp.scorechain.com"
tool_timeout_sec = 300

Privacy & Support

Don't paste API keys, customer credentials, or PII into chat prompts — they persist in the model's context window.
For access, billing, or production issues: support@scorechain.com

Scorechain MCP

Authentication​

API Key​

OAuth (sign in with Scorechain)​

Connect Your Client​

Claude Desktop​

Claude Code​

Cursor​

VS Code (GitHub Copilot Chat)​

ChatGPT​

OpenCode​

Other Clients​

Tools​

Resources​

Prompts​

How to Read Scores​

Example Prompts​

Connection Details​

Troubleshooting​

Privacy & Support​