> ## Documentation Index
> Fetch the complete documentation index at: https://pro-bot.dev/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Run your bot for free with Azure student credits

> Students get $100 in Azure credits - enough to deploy your own model and power a ProBot for up to a year. Here's the full walkthrough.

*A ProBot resource guide*

***

<p align="justify">
  ProBot is bring-your-own-key: you connect your own model, and you pay your
  provider directly (often pennies per conversation). But "pennies" still means
  a card on file - and if you're a student, you can skip that entirely.
</p>

<p align="justify">
  **[Azure for Students](https://azure.microsoft.com/en-us/free/students/) gives
  you \$100 in Azure credit with no credit card required.** That's more than
  enough to deploy your own small chat model and run a ProBot on it for months.
  This guide walks you from a fresh student account to a live bot, end to end.
</p>

<Note>
  This is the closest thing to a truly \$0 hosted bot: Azure covers the model
  bill from your free credit, and ProBot's hosted app is free to use. The one
  honest caveat - student credit lasts **12 months** (renewable while you're
  enrolled), so it's "free for now," not "free forever." We cover how to make it
  last at the end.
</Note>

## Why this works so well

<p align="justify">
  ProBot already speaks Azure OpenAI natively. Once you deploy a model in Azure,
  you get four values - an endpoint, a deployment name, an API key, and an API
  version - and ProBot has a field for each. A resume-Q\&A bot is light work for
  a "mini"-class model, so your \$100 stretches a long way.
</p>

| What you get from Azure | Where it goes in ProBot                           |
| ----------------------- | ------------------------------------------------- |
| Endpoint URL            | **Azure endpoint**                                |
| Deployment name         | **Deployment name** (used as the model)           |
| API key                 | **API key**                                       |
| API version             | **API version** (a sensible default is prefilled) |

## Before you start

<p align="justify">You'll need:</p>

* A school email address (or [GitHub Student](https://education.github.com/pack) verification).
* To be 18+ and a full-time student at an accredited institution.
* About 15 minutes. **No credit card.**

## Step 1 - Claim your \$100 Azure student credit

<Steps>
  <Step title="Open the Azure for Students offer">
    Go to
    [azure.microsoft.com/free/students](https://azure.microsoft.com/en-us/free/students/)
    and click **Start free**. Sign in with a personal Microsoft account (or
    create one).
  </Step>

  <Step title="Verify you're a student">
    Enter your details and your **school email**. Azure verifies your academic
    status automatically; if it can't, it'll ask you to upload a student ID or
    verify through GitHub Student. No card is requested at any point.
  </Step>

  <Step title="Land in the Azure portal">
    Once approved you'll have a subscription with **\$100 of credit** valid for
    12 months. You can see the remaining balance any time under **Subscriptions
    → your subscription**.
  </Step>
</Steps>

<Tip>
  The credit doesn't roll over, but you can **renew each year you're enrolled**
  for a fresh \$100. Set a calendar reminder a few weeks before it expires.
</Tip>

## Step 2 - Deploy a model

<p align="justify">
  Azure's AI tooling now lives in the **[Microsoft Foundry
  portal](https://ai.azure.com)** (you may also see it called Azure AI Foundry
  or Azure OpenAI). The flow is the same: create a resource, then deploy a model
  into it.
</p>

<Steps>
  <Step title="Create an Azure OpenAI resource">
    In the [Azure portal](https://portal.azure.com), search **Azure OpenAI** →
    **Create**. Pick your student subscription, a resource group (create one if
    needed), and a region, then create it. Microsoft's official walkthrough is
    [here](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/create-resource).
  </Step>

  <Step title="Open the Foundry portal and deploy a model">
    From the resource, open **Microsoft Foundry** (or go to
    [ai.azure.com](https://ai.azure.com)). Go to **Models + endpoints** →
    **Deploy model**, and choose an inexpensive chat model such as
    **`gpt-4o-mini`**. Give the deployment a name you'll remember - that name is
    what you'll use as the model in ProBot.
  </Step>

  <Step title="Wait for the deployment to finish">
    It takes a minute or two. When it's ready, it shows up in your **Models +
    endpoints** list.
  </Step>
</Steps>

<Tip>
  Choose a **mini / small** model. For answering questions about your resume you
  do not need a flagship model, and the cheaper one makes your \$100 last
  dramatically longer.
</Tip>

## Step 3 - Copy the four values ProBot needs

<p align="justify">
  Open your deployment's detail page (**Models + endpoints → your deployment**).
  From there, collect:
</p>

* **Endpoint** - the resource URL, e.g. `https://your-resource.openai.azure.com` (it must start with `https://`).
* **Deployment name** - exactly what you named the deployment in Step 2.
* **API key** - copy **Key 1** from the resource's **Keys and Endpoint** page. Treat it like a password.
* **API version** - e.g. `2025-01-01-preview`. ProBot prefills a working default, so you can usually leave it.

<Warning>
  Your API key is a secret. Don't paste it into chats, screenshots, or commits.
  In ProBot it's kept encrypted, sent only to Azure, and never logged in
  plaintext.
</Warning>

## Step 4 - Plug it into ProBot

<Steps>
  <Step title="Open AI model & key">
    In ProBot, go to **AI model & key** (sidebar → *Manage model & key*, or
    Settings → AI model & key) - or set it while building the bot in the Bot
    Factory.
  </Step>

  <Step title="Pick Azure and fill the fields">
    Choose **Azure** as the provider, then paste your **endpoint**, **deployment
    name** (this is the model), **API key**, and leave the **API version** at
    its default unless you have a reason to change it.
  </Step>

  <Step title="Save and test">
    Save, then open your bot and ask it a question. If it replies, you're live -
    running on your own model, paid for by your Azure credit.
  </Step>
</Steps>

<p align="justify">
  For the exact field-by-field reference, see [Models & keys → Azure
  specifics](/guides/models-and-keys).
</p>

## Make your \$100 last

<p align="justify">A few habits keep the credit going for the long haul:</p>

* **Set a spending alert.** In the Azure portal under **Cost Management → Budgets**, create a budget (say \$5/month) so you're warned long before the credit runs low.
* **Stick to a mini model.** It's the single biggest lever on cost.
* **Lean on ProBot's guardrails.** Per-bot rate limits and input sanitization are on by default, so a curious visitor can't burn through your credit.
* **Watch the expiry date.** Renew your student offer each year, or switch the bot to another provider's free tier before the credit lapses.

## A note on honesty

<p align="justify">
  This path is genuinely free *while your student credit lasts*. When it
  expires, your bot doesn't have to die - ProBot lets you swap providers any
  time (including a fully free local model via [Ollama](/guides/models-and-keys)
  if you self-host). Your bot, your model, your call.
</p>

***

*Ready to point a model at your bot? [Get started free](https://pro-bot.dev/register), and see [Models & keys](/guides/models-and-keys) for every supported provider.*