Welcome to the Itzam documentation! Here you will learn how to create AI features in your projects using Itzam.

​
How it works

Let’s use a real-world example to understand how Itzam works. Imagine you want to create an AI support chat.

​
Workflow

Everything starts in the dashboard, where you will create a Workflow. A Workflow represents a functionality in your app, in this case, the support chat.
  1. Create a Workflow called “Support Chat”. Create a Workflow
  2. Inside your Workflow you can configure:
    • AI model (GPT, Claude, Gemini)
    • Prompt (e.g. “You are a customer support agent from Acme Inc…”)
    • Model settings (response length, style)
    • Knowledge (what the model will use as context, e.g. docs, links)
    Workflow configuration
  3. Now create a new API key in the dashboard. Create an API key
  4. Great! Now grab this Workflow’s slug (in this case, support-chat) and your API key and let’s integrate!
You have two options to integrate this Workflow in your app:

​
SDKs

  1. Install the SDK:
npm i itzam
  1. Use the API key to authenticate:
import Itzam from "itzam";

const itzam = new Itzam("itzam_81d443ca-3952-4c53-bfda-7...");
  1. Use the Workflow’s slug to run it with your user’s input:
const response = await itzam.streamText({
  input: "Where is your documentation?",
  workflowSlug: "support-chat",
});

// response = "Follow this link to access our documentation: acme.com/docs"

​
API

If you prefer, you can also use the API directly. 
curl -X POST \
  -H "Api-Key: YOUR_API_KEY" \
  -d '{"input": "Hello, world!", "workflowSlug": "my_great_workflow"}' \
  https://itz.am/api/v1/stream/text

​
Nice work! 🎉

You have created your first Workflow and integrated it in your app. Now you can manage everything about it in the dashboard. You can:

​
Change the AI model

Change model

​
Update the model settings

Update model settings

​
Tweak the prompt

Tweak prompt

​
Add knowledge (docs, links, etc.)

Add knowledge

​
Check your runs

Check runs

​
Make tests in playground

Test inputs