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

# API & SDK Overview

> Access Bud AI Foundry programmatically via REST APIs, Python SDK, and CLI tools

## Introduction

Bud AI Foundry provides multiple ways to interact with the platform programmatically:

* **Python SDK**: High-level interface for Python applications
* **Gateway API**: OpenAI-compatible REST API for model inference
* **CLI Tool**: Command-line interface for automation and DevOps

## Quick Comparison

| Interface       | Best For                                    | Installation                                                       |
| --------------- | ------------------------------------------- | ------------------------------------------------------------------ |
| **Python SDK**  | Python applications, data science workflows | `pip install git+https://github.com/BudEcosystem/BudAIFoundry-SDK` |
| **Gateway API** | Any language, direct HTTP access            | REST API calls                                                     |
| **CLI**         | DevOps, CI/CD, shell scripts                | `pip install git+https://github.com/BudEcosystem/BudAIFoundry-SDK` |
| **OpenAI SDK**  | Drop-in OpenAI replacement                  | Change base URL                                                    |

## Key Capabilities

### Model Inference

```python theme={null}
from budai import BudClient

client = BudClient(api_key="your-key")
response = client.chat.completions.create(
    model="llama-3.2-1b",
    messages=[{"role": "user", "content": "Hello!"}]
)
```

### Pipeline Automation

```python theme={null}
# Create and execute deployment pipelines
pipeline = client.pipelines.create(
    name="Deploy Model",
    definition={
        "steps": [...]
    }
)
```

### Infrastructure Management

```python theme={null}
# Manage clusters and deployments
cluster = client.clusters.get("cluster_prod")
deployment = client.deployments.create(
    model_id="model_xyz",
    cluster_id=cluster.id
)
```

## Authentication

All API access requires authentication:

1. **API Keys**: Generate in the dashboard under API Keys & Security
2. **Environment Variable**: Set `BUD_API_KEY` for SDK/CLI
3. **Header**: Use `Authorization: Bearer <key>` for direct API calls

See [Authentication](/api-sdk/python-sdk/authentication) for detailed setup.

## OpenAI Compatibility

The Gateway API is fully compatible with OpenAI's interface:

```python theme={null}
from openai import OpenAI

# Just change the base URL
client = OpenAI(
    base_url="https://gateway.bud.studio/v1",
    api_key="your-bud-key"
)

# Use OpenAI SDK as normal
response = client.chat.completions.create(
    model="llama-3.2-1b",
    messages=[{"role": "user", "content": "Hello!"}]
)
```

## Support

* **Documentation**: Browse the sections in this tab
* **GitHub**: [BudAIFoundry-SDK](https://github.com/BudEcosystem/BudAIFoundry-SDK)
* **Issues**: Report bugs and request features on GitHub
* **Email**: [contact@bud.studio](mailto:contact@bud.studio) for enterprise support

## Next Steps

<CardGroup cols={2}>
  <Card title="Quickstart" icon="play" href="/api-sdk/quickstart">
    Make your first API call in 5 minutes
  </Card>

  <Card title="Python SDK" icon="python" href="/api-sdk/python-sdk/installation">
    Install and use the Python SDK
  </Card>

  <Card title="Gateway API" icon="plug" href="/api-sdk/chat/create-chat-completion">
    OpenAI-compatible inference API
  </Card>

  <Card title="CLI Reference" icon="terminal" href="/api-sdk/python-sdk/cli-reference">
    Command-line tools for DevOps
  </Card>
</CardGroup>