1. Create your first agent. Sidebar → Email Campaigns → Agents → New agent. Pick a persona, language, and audience filter.
2. Import or discover prospects. Use Prospects → Import CSV, or ask PhewDo: "Find 200 fintech founders in UAE".
3. Write a drip. Open Drip sequences, click New sequence. PhewDo can draft all 4 steps for you.
4. Connect a sending domain. AI Inbox → Setup → Domain. We auto-issue SPF/DKIM and warm it for 14 days.
5. Launch. Toggle the agent ON. Sends start in the next 5-minute sweeper window.

Agents run independently. Each agent owns: a persona, a target ICP, its own drip sequences, a sending mailbox, and an inbox view. Reply detection pauses sends per-prospect automatically.

Three ways to add prospects: CSV import (with field mapping), discovery (LinkedIn + web search), or via the API. Every prospect is enriched (verified email, mobile, role, company size) before it enters a drip.

• Tags & segments for any-size list
• Bounce-protection: invalid emails auto-suppressed
• Per-agent prospect pool, no overlap

A drip is a sequence of 1–10 steps with configurable delays. Quiet hours (default 10pm–8am IST) are enforced globally. Variants enable automatic A/B winners.

curl https://api.phewdo.com/phewdo-api \
  -H "Authorization: Bearer phw_live_..." \
  -H "Content-Type: application/json" \
  -d '{"action":"get_agents"}'

const res = await fetch('https://api.phewdo.com/phewdo-api', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${process.env.PHEWDO_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({ action: 'get_agents' }),
});
const { agents } = await res.json();
console.log(agents);

import os, requests

resp = requests.post(
    'https://api.phewdo.com/phewdo-api',
    headers={
        'Authorization': f"Bearer {os.environ['PHEWDO_KEY']}",
        'Content-Type': 'application/json',
    },
    json={'action': 'get_agents'},
)
print(resp.json()['agents'])

Every request is a POST to /phewdo-api with a JSON body of the form {"action":"…", …params}. Replies are JSON {"ok":true, …}. The same dispatcher serves the dashboard — anything you can do in the UI maps to one action.

What you can do with this key

The key (and the MCP server backed by it) lets your AI or your scripts do anything a logged-in user can do in the dashboard — read data, push prospects, build drips, send emails, manage audience, draft replies. Everything that mutates billing, credentials, plan, or platform settings is locked away.

Rule of thumb — if the action would cost money or change your account's identity, it's locked to JWT (i.e. you logging into the dashboard). Everything else is fair game for keys.

Full action catalogue

Response codes

The MCP server speaks the standard Model Context Protocol (Streamable HTTP). Authenticated with your API key — same one you use for REST. Connect once, your AI gets PhewDo as native tools.

Connect from Claude Desktop / Claude.ai / Cursor

{
  "mcpServers": {
    "phewdo": {
      "url": "https://mcp.phewdo.com/mcp",
      "headers": {
        "Authorization": "Bearer phw_live_..."
      }
    }
  }
}

Step-by-step setup

Pick your client and follow the 3 steps. The same key works in all of them.

{
  "mcpServers": {
    "phewdo": {
      "url": "https://mcp.phewdo.com/mcp",
      "headers": {
        "Authorization": "Bearer phw_live_..."
      }
    }
  }
}

Tools your AI gets (15 total)

mcp.phewdo.com/tools.json returns the live tool list (no auth needed — handy for sanity-checking connectivity).

Key rotation & revocation

• Lose a key? Revoke it from this page (any active key has a button). Revocation is instant — the next call returns 401.
• You can hold multiple keys at once — one per integration (Claude Desktop, n8n, scripts) makes audit trails easier.
• Keys are stored hashed with bcrypt — even we can't read your token back. The plaintext is shown once at creation. Save it in a password manager.
• Plan-gating is enforced on every call — if you downgrade below AI Inbox Business your keys start returning 402 until you upgrade again.

Add a webhook URL in Profile → Webhooks. We POST a signed JSON payload on every event. Retries use exponential backoff for 24 hours.

Event types

• 60 requests / minute per API key
• 5,000 requests / day per API key (rolling 24h window)
• 1,000 prospects per bulk-insert call

Every response carries: X-RateLimit-Limit-Minute, X-RateLimit-Limit-Day, X-RateLimit-Remaining-Minute, X-RateLimit-Remaining-Day. When you exceed a window you get 429 with a Retry-After header (seconds) and JSON body {"ok":false,"retry_after_seconds":N}.

Need more? Enterprise plans get custom limits — email hello@phewdo.com.

Connect from Modules → CRM. Maps prospect fields, mirrors deal stages, and pushes replies as notes. Supports webhook triggers + bulk sync.

• Bidirectional prospect ↔ contact sync
• Custom field mapping with type validation
• Stage-based automations (e.g. "moved to Qualified → start drip")

• Auto-configured authentication on every domain
• Per-mailbox daily caps, throttling, quiet hours
• Open + click + bounce tracking via signed webhook events
• Reputation monitoring with daily health score

Search by company size, industry, geography, role, tech stack, or natural-language ICP. Results are deduped, enriched (email, mobile, LinkedIn URL), and verified before they hit your pipeline.

• Per-agent booking pages with custom availability
• Auto-generated video link on confirmation
• Live transcription & speaker diarization
• Post-call AI summary, action items, CRM-ready notes

Same drip engine, different rails. Per-region sender pools, template approval workflow, link tracking on shortened URLs.

• 3-day free trial · no card required
• Upgrade/downgrade prorated to the day
• Add-ons: domains, mailboxes, drip packs, meeting booker, AI replies
• Invoices auto-issued; tax fields configurable in Profile → Billing

• TLS 1.3 everywhere, AES-256 at rest
• Per-workspace row-level isolation
• SOC 2 Type II in progress · GDPR & DPDPA compliant
• SSO (SAML/OIDC) on Enterprise
• Audit log of every config change & API call