Build with AI (Claude, ChatGPT & friends)

If you're building your product with Claude, ChatGPT, Cursor or any other AI assistant, you don't need to read our docs — your assistant does, and we've made that trivial.

The two URLs your assistant should know

• https://wezend.com/llms.txt — a compact map of everything on this site (llms.txt convention).
• https://wezend.com/llms-full.txt — the entire documentation as one plain-text file: every endpoint, every field, every example. Small enough for a modern context window, complete enough to build against without browsing.

If your assistant can fetch URLs, one instruction suffices: "Read https://wezend.com/llms-full.txt and integrate WeZend."

The copy-paste context pack

No URL fetching? Paste this block into the conversation instead — it's the minimum an LLM needs to integrate correctly:

WEZEND API — INTEGRATION CONTEXT
Base URL: https://api.wezend.com  (all paths below are under /v1)
Auth: header  X-API-Key: <key>   (server-side only — NEVER in browser code)

SEND A MESSAGE (any channel: sms | email | whatsapp | voice | push | inapp)
  POST /v1/messages/send
  { "to": "+4512345678" | "a@b.com", "channel": "sms", "sender": "Acme",
    "message": "text", "subject": "(email)", "html": "(email)",
    "template_id": "tpl_… (optional)", "variables": { }, "webhook_url": "(optional)" }
  → 201 { "message_id", "status": "queued", "cost", "currency" }
  Status: GET /v1/messages/:id   Bulk: POST /v1/messages/bulk { "messages": [ … ] }

CONTACTS (CDP)
  POST  /v1/contacts             { "email", "phone", "first_name", …custom fields }
  PATCH /v1/contacts/:id         (partial update; matched/deduped on email/phone)
  POST  /v1/contacts/fields      { "key", "type": "text|number|date|boolean", "label" }
  POST  /v1/contacts/import      (CSV)

BEHAVIOURAL EVENTS (feeds segments & journeys)
  POST /v1/events
  { "contact": { "email": "a@b.com" }, "event": "order_completed",
    "properties": { "value": 499, "currency": "DKK" } }

LEAD FORMS (public — safe in browser code, no API key)
  POST https://api.wezend.com/forms/<form-key>/submit   { "email", …fields }
  Hosted page: GET https://api.wezend.com/forms/<form-key>

DELIVERY WEBHOOKS (configure URL in dashboard or per-send webhook_url)
  You receive: POST { "message_id", "status": "delivered|failed|…", "timestamp", … }
  Verify the X-WeZend-Signature header with your webhook secret (GET /v1/webhook-secret).

RULES FOR GENERATED CODE
  1. API key lives in an env var (WEZEND_API_KEY), called from backend code only.
  2. Public browser code may only call /forms/<key>/submit — nothing else.
  3. Marketing sends require consent; transactional sends set "category": "transactional".
  4. Use E.164 phone numbers (+4512345678). Handle 402 (insufficient balance) and 429 (rate limit).
Full reference: https://wezend.com/llms-full.txt

Prompts that work

"Add a newsletter signup to my site's footer that submits to my WeZend form site-footer with a success state."

"When a user completes checkout in this Express app, send an order-confirmation email via WeZend and track an order_completed event with the order value."

"Build a /webhooks/wezend endpoint that verifies the signature and marks messages delivered in my DB."

The assistant has everything it needs in the pack above — endpoints, shapes, and the security rules that keep it from putting your API key in a browser bundle.

Getting your credentials

1. Create an account and grab a key under Settings → API keys.
2. Create a form under Audience → Forms if the assistant is building signup UI.
3. Set WEZEND_API_KEY in your project's environment — and let the assistant do the rest.

Want to operate, not just build?

With Connect AI (MCP) your assistant doesn't just write code against WeZend — it runs your account directly: 34 customer-scoped tools for segments, campaigns, journeys, domains and budgets, with real sends behind an explicit confirmation.