Wire DataVibe into your AI SDR in 15 minutes

Use this guide when you have an AI agent that generates outbound email, support replies, or workflow actions. DataVibe checks the output **before** it reaches a customer.

When to use which API

Production integrations should use **POST /v1/gate/outbound** so submissions land in the approval queue with a full audit trail.

1. Get your API key

1. Sign up at app.datavibe.cc/signup?program=alpha 2. Open Alpha Quickstart and copy your workspace API key 3. Store it as DATAVIBE_API_KEY in your agent environment

2. Apply the AI SDR policy pack

In the dashboard:

1. Go to **Onboarding → Setup wizard** 2. Choose use case **AI SDR / outbound** 3. Apply template **ai_sdr_guardrails_v1**

This enables rules for pricing hallucination, false guarantees, competitor mentions, and tone violations.

3. Minimal integration (Node.js)

const DATAVIBE_API_KEY = process.env.DATAVIBE_API_KEY;
const CORE_API = process.env.DATAVIBE_BASE_URL ?? "https://api.datavibe.cc";

async function gateBeforeSend({ recipient, subject, bodyText }) {
  const res = await fetch(`${CORE_API}/v1/gate/outbound`, {
    method: "POST",
    headers: {
      Authorization: `Bearer ${DATAVIBE_API_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      recipient,
      subject,
      body_text: bodyText,
      body_html: `<p>${bodyText.replace(/\n/g, "<br/>")}</p>`,
      source_model: "your-llm",
      metadata: { integration: "ai_sdr" },
    }),
  });

  const data = await res.json();
  if (data.status === "BLOCKED") {
    throw new Error(`Blocked by policy: ${JSON.stringify(data.policy_violations)}`);
  }
  if (data.status === "QUEUED") {
    // Human review required — do not send until approved in dashboard
    return { send: false, actionId: data.action_id, reviewUrl: data.review_url };
  }
  return { send: true, actionId: data.action_id };
}

// Usage in your agent loop:
// const aiDraft = await llm.generateOutboundEmail(lead);
// const gate = await gateBeforeSend({ recipient: lead.email, subject: aiDraft.subject, bodyText: aiDraft.body });
// if (gate.send) await sequencer.send(...);

4. Python (SDK)

from datavibe import DataVibeClient

client = DataVibeClient(api_key="dv_live_...")

result = await client.check(
    content=ai_generated_text,
    content_type="email",
)

if result.verdict == "blocked":
    log_blocked(result.violations)
elif result.verdict == "review_required":
    queue_for_human_review(result)
else:
    send_email(ai_generated_text)

5. HubSpot / Zapier pattern

**Generate → Gate → Send only if safe**

1. **Trigger:** New lead or AI draft ready 2. **Action:** HTTP POST to https://api.datavibe.cc/v1/gate/outbound with the AI-generated subject and body 3. **Branch:** - status === "BLOCKED" → alert RevOps, do not send - status === "QUEUED" → notify reviewer (Slack/email), wait for approval in Approval Queue - status === "SENT" or approved → continue to email/CRM step

Optional: poll GET /v1/gate/outbound/:action_id or use dashboard webhooks for approval events.

6. Test payloads (90-second demo)

Run these in order from Alpha Quickstart:

Confirm results in Approval Queue and Audit Log.

7. Next steps

• Full integration guide
• Live playground — try without an account
• AI SDR policy simulator
• Founding Alpha pricing — $99/mo, 10,000 checks

Questions: [email protected]