Decision Intelligence MCP server
{
"mcpServers": {
"ontoramp-decision-intelligence": {
"type": "http",
"url": "https://ontoramp-decision-intelligence.fly.dev/mcp"
}
}
}{
"mcpServers": {
"ontoramp-decision-intelligence": {
"type": "http",
"url": "https://ontoramp-decision-intelligence.fly.dev/mcp"
}
}
}{
"mcpServers": {
"ontoramp-decision-intelligence": {
"type": "http",
"url": "https://ontoramp-decision-intelligence.fly.dev/mcp"
}
}
}{
"servers": {
"ontoramp-decision-intelligence": {
"type": "http",
"url": "https://ontoramp-decision-intelligence.fly.dev/mcp"
}
}
}{
"mcpServers": {
"ontoramp-decision-intelligence": {
"type": "http",
"url": "https://ontoramp-decision-intelligence.fly.dev/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.0.1
What it does
OntoRamp Decision Intelligence logs, evaluates, and validates AI-assisted decisions against an organization's authority graph, returning PASS, WARN, or BLOCK outcomes. It is for teams building agentic systems who need governance, audit trails, and precedent lookup for consequential decisions.
- Look up prior decisions and precedents before an AI agent takes a consequential action.
- Log a decision with full context and receive an evaluation of authority gaps and governance alignment.
- Build an audit trail of AI-assisted decisions for compliance and governance review.
Tools this server exposes
| Tool | Does |
|---|---|
| find_decision_precedent | Look up prior decisions relevant to a proposed action (read-only, available on Free tier). |
| log_decision | Record a consequential AI-assisted decision with authority context (requires Developer+ tier and authentication). |
| evaluate_decision | Evaluate a decision for authority gaps, precedent conflicts, and governance alignment (requires Developer+ tier). |
| request_assessment | Submit an Agentic Readiness Assessment request (available on all tiers, rate-limit exempt). |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.