> For the complete documentation index, see [llms.txt](https://docs.collieai.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.collieai.io/getting-started/overview.md).

# Getting started with CollieAi

How to get started with CollieAi, an AI firewall and LLM security gateway — secure your LLM app, block prompt injection, and redact PII in about five minutes.

CollieAi is an AI firewall (LLM security gateway) that sits between your application and your LLM provider. Every request and response passes through a configurable filtering pipeline that detects and redacts PII, blocks prompt injection and jailbreaks, filters URLs, and more — without changing how your application talks to the model.

{% hint style="info" %}
**Key points**

* CollieAi is a drop-in, provider-agnostic proxy — switch by changing one `base_url`. It works with any major LLM provider through an OpenAI-compatible API and Anthropic's native Messages API.
* It works with OpenAI and Anthropic Claude out of the box, and with any OpenAI-compatible endpoint (including self-hosted models like vLLM, Ollama, or LocalAI) via a custom base URL. Native Gemini, Azure OpenAI, and Bedrock support is on the roadmap.
* Two integration paths: a real-time drop-in proxy, and async jobs for batch or long-running work.
* You need a CollieAi account and a provider API key to send your first secured request.
  {% endhint %}

<figure><img src="https://4140227778-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxKzkxBScfGXAbqoqyEbd%2Fuploads%2FCM4tJmhEg9m8wD5GR4fA%2F1.png?alt=media&amp;token=0da8ddcd-d118-42b3-b192-2567d3dbf235" alt=""><figcaption></figcaption></figure>

## What do you need to get started?

Before you begin, make sure you have:

* [x] **A CollieAi account** - Sign up at [app.collieai.io](https://app.collieai.io/) using Google OAuth or email/password.
* [x] **A provider API key** - You need an API key from your LLM provider — OpenAI or Anthropic (or any OpenAI-compatible endpoint). CollieAi proxies your requests using this key.

## Three Ways to Integrate

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th></tr></thead><tbody><tr><td><h4><i class="fa-right-to-bracket" style="color:$primary;">:right-to-bracket:</i></h4></td><td><a href="/proxy-integration/overview.md"><strong>Drop-in Proxy</strong></a></td><td>The fastest path. Point your OpenAI or Anthropic SDK (or any OpenAI-compatible client) at CollieAi by changing the <code>base_url</code>.<br><br>Your existing code works as-is - CollieAi transparently applies security rules and forwards requests to the provider.<br><br>Best for: real-time chat, streaming responses, low-latency use cases.</td></tr><tr><td><h4><i class="fa-rotate" style="color:$primary;">:rotate:</i></h4></td><td><a href="/async-jobs/overview.md"><strong>Async Jobs</strong></a></td><td>Submit a request and receive a callback when the result is ready.<br><br>Useful for batch processing, long-running prompts, or workflows where you do not need an immediate response.<br><br>Best for: background processing, batch pipelines, webhook-driven architectures.</td></tr><tr><td><h4><i class="fa-code" style="color:$primary;">:code:</i></h4></td><td><a href="/sdks/overview.md"><strong>SDKs</strong></a></td><td>Language-native clients for Python, Node, and .NET. Run your own model and stream back only CollieAi-released text - the SDK owns the chunk protocol and input/output moderation, so you never touch the raw endpoints.<br><br>Best for: customer-owned streaming, safe token-by-token UX from your own backend.</td></tr></tbody></table>