Zapier + Ertas

Connect Ertas-trained models to 6,000+ apps through Zapier's no-code automation platform for intelligent document processing, email triage, and workflow automation.

Overview

Zapier is the world's most widely used no-code automation platform, connecting over 6,000 applications through event-driven workflows called Zaps. When something happens in one app — a new email arrives, a form is submitted, a file is uploaded — Zapier automatically triggers actions in other apps: creating records, sending notifications, updating spreadsheets, or calling APIs. With millions of workflows running daily, Zapier has become the connective tissue of modern business operations.

Zapier's AI integration capabilities have expanded rapidly, with native support for calling language model APIs as workflow steps. This means any business process that involves text understanding — classifying support tickets, extracting data from emails, summarizing meeting notes, or generating responses — can now incorporate AI intelligence. The key limitation of generic AI in Zapier workflows is accuracy: a general-purpose model may misclassify industry-specific content, miss domain jargon, or generate responses that do not match your organization's voice. This is exactly the problem that Ertas-trained models solve.

How Ertas Integrates

Ertas-trained models connect to Zapier through the Webhooks by Zapier action or the Code by Zapier step, both of which can call any REST API. After fine-tuning a model in Ertas Studio and deploying it to Ertas Cloud or any OpenAI-compatible endpoint, you add an API call step in your Zap that sends text to your model and receives the AI-generated response. Because the API contract follows the OpenAI format, the integration requires minimal configuration — just the endpoint URL, API key, and the message payload.

The power of this integration lies in the combination of Zapier's breadth of app connections with Ertas's depth of domain-specific AI. A law firm can build a Zap that automatically classifies incoming emails by case type, extracts key dates and party names, and routes them to the appropriate attorney — using a model fine-tuned on their specific practice areas. An ecommerce company can auto-generate product descriptions from supplier data sheets, classify customer reviews by sentiment and topic, and draft personalized response emails — all running as automated Zaps without human intervention. The fine-tuned model ensures these automations are accurate enough for production use, not just impressive demos.

Getting Started

1
Fine-tune your model in Ertas Studio
Train a model on your specific automation task — email classification, data extraction, content generation, or any text processing workflow you want to automate.
2
Deploy to a publicly accessible endpoint
Serve your model via Ertas Cloud or any inference server with a public URL. Zapier needs to reach the endpoint over the internet.
3
Create a Zap with your trigger
Set up the trigger event in Zapier — new email, form submission, spreadsheet row, or any of the 6,000+ supported app events.
4
Add a Webhooks action to call your model
Add a Webhooks by Zapier POST action pointing to your model's /v1/chat/completions endpoint. Configure the request body with the trigger data as the user message.
5
Route the AI response to downstream actions
Use the model's response in subsequent Zap steps — update a CRM record, send a notification, create a task, or trigger another workflow based on the AI classification.

json

{
  "action": "Webhooks by Zapier - POST",
  "url": "https://cloud.ertas.ai/v1/chat/completions",
  "headers": {
    "Content-Type": "application/json",
    "Authorization": "Bearer {{ertas_api_key}}"
  },
  "body": {
    "model": "ertas-support-classifier-7b",
    "messages": [
      {
        "role": "system",
        "content": "Classify the support ticket into: billing, technical, feature-request, or escalation. Return only the category."
      },
      {
        "role": "user",
        "content": "{{trigger.email_body}}"
      }
    ],
    "temperature": 0.0,
    "max_tokens": 20
  }
}

Zapier Webhooks action calling an Ertas-trained model to classify incoming support tickets by category.