Developer Platform

Build Agents. Sell Them. Earn Money.

Create governed AI agents, publish them to the DingDawg marketplace, and earn revenue every time a business installs your agent.

Start BuildingAPI Reference
Quick Start

Three steps to your first agent

From zero to published in under 30 minutes. Every agent runs on the governed runtime — compliance, receipts, and billing handled automatically.

1

Create

Define your agent with a manifest and capabilities

manifest.json
{
  "name": "@yourname/no-show-recovery",
  "version": "1.0.0",
  "description": "Recovers missed appointments via SMS + email",
  "author": {
    "name": "Your Name",
    "email": "you@example.com",
    "url": "https://yoursite.com"
  },
  "capabilities": ["sms.send", "email.send", "booking.read"],
  "pricing": {
    "model": "per_action",
    "price": 0.50,
    "currency": "usd"
  },
  "entry": "./src/index.ts",
  "governance": "./governance/policies.yaml"
}
2

Govern

Gate every action before it executes — get a signed receipt

test.ts
// govern_action returns a signed receipt with risk score + decision

const result = await client.callTool("govern_action", {
  agent_id: "my-agent",
  action_type: "send_email",
  action_description: "Sending weekly report to 500 subscribers",
  target_resource: "email_server",
  risk_tier: "medium",
});

console.log(result.decision);          // "allow" | "review" | "deny"
console.log(result.risk_score);        // 0–100 (0–39 allow, 40–69 review, 70+ deny)
console.log(result.receipt_id);        // "gov_1x2y3z_ab12cd34"
console.log(result.policy_violations); // [] or list of violations
console.log(result.recommended_controls); // ["Validate content before sending"]
3

Install

Add to Claude Code, Cursor, or any MCP client in one command

terminal
# Add to Claude Code (1 command)
claude mcp add dingdawg-governance npx dingdawg-governance

# Verify it loaded
claude mcp list

# Or add to Cursor / any MCP client — ~/.cursor/mcp.json:
# {
#   "mcpServers": {
#     "dingdawg-governance": {
#       "command": "npx",
#       "args": ["dingdawg-governance"]
#     }
#   }
# }

# Receipts stored locally at:
# ~/.dingdawg/governance/receipts/<receipt_id>.json

# With API key — receipts sync to cloud, team audit trail, unlimited checks:
# DINGDAWG_API_KEY=your-key npx dingdawg-governance
SDK

Wire governance into your agent

Use the dingdawg-governance MCP server to gate every agent action with a signed receipt.

governed-agent.ts
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js";

const transport = new StdioClientTransport({
  command: "npx",
  args: ["dingdawg-governance"],
});

const client = new Client({ name: "billing-agent", version: "1.0.0" }, {});
await client.connect(transport);

// Govern an action before it executes
const receipt = await client.callTool("govern_action", {
  agent_id: "billing-agent",
  action_type: "make_purchase",
  action_description: "Purchasing $450 cloud credits for infrastructure renewal",
  target_resource: "payment_api",
  risk_tier: "high",
});

if (receipt.decision === "deny") throw new Error("Blocked: " + receipt.policy_violations[0].description);
if (receipt.decision === "review") { await notifyHuman(receipt.receipt_id); return; }

// Execute with governance receipt attached
await executePurchase({ receipt_id: receipt.receipt_id });
Revenue

You set the price. You earn the revenue.

DingDawg takes a 20% platform fee. You keep 80% of every action your agent performs.

Per Action

$0.01 - $5.00

Charge per action your agent performs. Best for utility agents.

Monthly Subscription

$9.99 - $499/mo

Recurring monthly fee for access. Best for full-service agents.

Free + Upgrade

$0 base

Free tier with paid upgrades. Best for growing your install base.

Everything you need to ship

Full REST API

Create, update, test, and publish agents programmatically.

Sandbox Environment

Isolated test environment with simulated payments and zero production risk.

Governed Runtime

Every action passes permission checks, context loading, and audit logging.

Neurosymbolic Governance

Your agent inherits the symbolic rules engine automatically. DingDawg Verified badge + IPFS proof on every governed call — zero compliance code to write.

Revenue Dashboard

Real-time earnings, install counts, action volume, and payout tracking.

MCP Server

One command to install. npx dingdawg-governance — works with Claude Code, Cursor, and any MCP-compatible agent.

Enterprise SDK — Custom Neurosymbolic Rules

Encode Your Exact Compliance Policies

Enterprise plan unlocks custom neurosymbolic ontologies — encode your organization's exact policies, industry-specific rules, and internal governance standards as immutable symbolic logic. Every agent call is verified against your rules, stored on IPFS, forever.

  • Custom compliance ontologies (Prolog-style rules)
  • Private IPFS cluster for audit proofs
  • Custom DingDawg Verified badge + your branding
  • Dedicated compliance reviewer
Talk to Enterprise Sales →

API Reference

Full endpoint documentation, auth, rate limits, and examples.

View docs

Pricing

Platform fees, payout schedule, and developer tier pricing.

View pricing

Package Format

Agent package spec, manifest schema, and publishing workflow.

View spec

Open Source

governance-sdk on GitHub — MIT licensed, full MCP server source with LNN interpretability.

View on GitHub

Loop Protocol

dingdawg-loop — safe scheduled proactive agents. Every loop execution passes a governance gate before acting.

npm install

Integration Guides

CrewAI, LangGraph, Claude Code, Cursor — governed by default in 60 seconds. Copy-paste snippets for every framework.

View guides

Start building today

Create a free developer account, build your first agent in sandbox, and publish to the marketplace when ready. 100 free API calls per day included.

Create Developer AccountRead the API Docs