Skip to content
For agents

Verifiable peg state in your prompts.

Pegana exposes peg-risk signals to AI agents over Model Context Protocol and REST. Subscribe to live state transitions, verify any alert's provenance, and cite receipts in the decision log. No API keys. Every tool is free.

methodology v0.7.0receipt v1sigstorefree

Agents that act on financial signals must cite why they acted.

Every Pegana signal carries an evidence_url and a receipt_sha256: a citable, verifiable record of every decision input, reproducible offline via pegana-replay. Severe transitions can additionally carry an on-chain anchor, a Solana SPL Memo commitment.

verify_alert.ts
const receipt = await mcp.callTool("verify_alert", {
  alert_id: alertUuid,
});

const { evidence } = receipt.structuredContent;
console.log("methodology:", evidence.methodology_version);
console.log("receipt sha256:", evidence.receipt_sha256);
console.log("on-chain:", evidence.onchain_tx_sig ?? "pending");

Every MCP tool is free. No keys, no signup, no payment. Rate-limited per IP so one agent cannot starve the rest.


Point your client here.

There's no browser UI. The server speaks JSON-RPC over POST; copy the endpoint into your MCP client.

mcp endpoint
https://mcp.pegana.xyz

# liveness  GET https://mcp.pegana.xyz/health
# metrics   GET https://mcp.pegana.xyz/metrics

Pegana is a verifiable computation log. Not SOC 2, not a regulatory audit, not financial advice.