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

> ## Agent Instructions
> Anchor Browser provides cloud browsers for AI agents and automation: stealth browsing with residential proxies, managed authentication into third-party web apps, and browser sessions that run reliably at scale. AI agents that need Anchor credentials should start with https://docs.anchorbrowser.io/quickstart/agent-access. This flow lets agents obtain an API key programmatically without creating a dashboard account. Read that page before attempting authentication or API access.

# Bring Your Own Model (BYOM)

> Connect your own LLM provider to power AI-driven browser automations at the project level

Bring Your Own Model (BYOM) lets you connect your own LLM provider — such as OpenAI, Anthropic, Google, or any OpenAI-compatible endpoint — to power AI-driven browser automations in Anchor. BYOM is configured at the **project level**: each project has its own model configuration, so different projects can use different providers and models.

**Why use BYOM?**

* **Cost control** — use your existing API agreements and volume discounts
* **Model choice** — pick the exact model that works best for each project's use case
* **Data privacy** — LLM requests go directly to your provider, never through Anchor's inference layer
* **Rate limits** — manage your own LLM rate limits

## Supported Providers

| Provider | Credentials | Example Models |
| - | - | - |
| OpenAI | `openai` | gpt-4o, gpt-4o-mini, o3 |
| Anthropic | `anthropic` | claude-sonnet-4-6, claude-haiku-4-5 |
| Google | `google` | gemini-2.5-pro, gemini-2.5-flash |
| Google Vertex AI | `vertex` | gemini-2.5-pro, gemini-2.5-flash |
| Azure OpenAI | `azure` | Any deployed Azure OpenAI model |
| Custom / OpenAI-compatible | `custom` | Any endpoint that follows the OpenAI chat completions API |

## Setup

<Steps>
  <Step title="Open your project's model settings">
    1. Navigate to your project in the Dashboard
    2. Go to **Project Settings → Model Configuration**
  </Step>

  <Step title="Add your API key">
    1. Click **Add Provider**
    2. Select your provider from the dropdown
    3. Paste your API key
    4. (Optional) Set a custom base URL — required for Azure and custom endpoints
    5. Click **Verify & Save** — Anchor will make a test call to confirm the key works
  </Step>

  <Step title="Choose a model">
    Select the specific model you want this project to use. Anchor will auto-populate available models for OpenAI, Anthropic, and Google. For custom endpoints, type the model name manually.
  </Step>

  <Step title="Save">
    Click **Save** to apply. All sessions created under this project will now use your configured model.

    Repeat these steps for each project that needs its own model. Projects without a BYOM configuration will continue using Anchor's default model.
  </Step>
</Steps>

## Using BYOM via the REST API

Include the `model` object in your `POST /projects` request:

```json theme={null}
{
  "model": {
    "provider": "openai",
    "name": "gpt-4o",
    "credentials": "sk-...",
    "base_url": null
  }

```

## FAQs

<AccordionGroup>
  <Accordion title="Does BYOM affect pricing?">
    Yes. When you bring your own model, Anchor does not charge for LLM usage — you pay your provider directly. You still pay for other Anchor browser usage metrics as usual.
  </Accordion>

  <Accordion title="Can I use different models for different projects?">
    Yes — that's the core idea. Each project has its own BYOM configuration. You can use GPT-4o in one project, Claude in another, and Anchor's default in a third.
  </Accordion>

  <Accordion title="Is my API key stored securely?">
    Project-level keys are encrypted. Session-level keys are never persisted — they exist in memory only for the session's lifetime.
  </Accordion>
</AccordionGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.