Your AI agents.
Full visibility. Total control.
One API call. Full audit trail. Human approval when it matters.
Send $15,000 to supplier
Visibility or control.
You choose.
Start with observe mode to see everything. Switch to enforce when you need approval gates.
Observe Mode
Track every action without blocking. Full visibility into what your AI agents are doing. Perfect for development and testing.
Enforce Mode
Require human approval for critical actions. AI proposes, you approve or reject. Webhook fires when you decide.
Every decision. Fully documented.
Automatic evidence collection for every action your agents take. Know exactly what happened, which policies applied, and why.
Agent, action type, parameters, timestamps - everything logged automatically.
See which policies matched, what rules triggered, and the final decision.
Who approved or rejected, when, and any notes they added.
Get started in 5 minutes
Add governance to your AI agents with just a few lines of code
Install the SDK
Choose your language and install
npm install apaai-ts-sdk
# or
pnpm add apaai-ts-sdkRegister Your Agent
Create and configure your AI agent
import { AccountabilityLayer } from 'apaai-ts-sdk';
const client = new AccountabilityLayer({
endpoint: 'https://api.apaai.ai',
apiKey: process.env.APAAI_API_KEY
});
// Register your agent
const agent = await client.agents.register({
name: 'marketing-bot',
description: 'Handles customer outreach and email campaigns',
config: {
webhookUrl: 'https://your-app.com/webhook'
}
});
console.log('Agent registered:', agent.id);Track Actions
Submit actions for governance evaluation
// Track an action - policies are evaluated automatically
const decision = await client.actions.create({
type: 'send_email',
agentId: agent.id,
actor: { kind: 'agent', name: 'marketing-bot' },
params: {
subject: 'Welcome!',
body: 'Thanks for signing up',
recipient: 'user@example.com'
}
});
// Check the governance decision
if (decision.status === 'approved') {
console.log('Action approved! Proceeding...');
await sendEmail(decision.params);
} else if (decision.status === 'pending_approval') {
console.log('Waiting for human approval...');
} else {
console.log('Action blocked by policy');
}Action passes all policies. Execute immediately.
Needs human review. Webhook sent.
Blocked by policy. Do not execute.
View in Dashboard
Monitor and approve actions in real-time
- View all action history with full audit trail
- Approve or reject pending actions with one click
- See which policies matched and why
- Real-time updates via webhooks or SSE
Simple, transparent pricing
Start free, scale as you grow. No hidden fees.
Starter
For individuals and small projects
Pro
For growing teams
Enterprise
For large organizations
All plans include 14-day free trial. No credit card required.