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

# Salad AI Gateway Overview

> OpenAI-compatible LLM API powered by SaladCloud distributed GPU infrastructure.

*Last Updated: April 09, 2026*

<Warning>
  Salad AI Gateway is currently in **closed beta**. To request access, sign up at
  [salad.com/ai-gateway](https://salad.com/ai-gateway).
</Warning>

## What is Salad AI Gateway?

Salad AI Gateway is a managed LLM API service that gives you direct access to powerful open-source models through a
single, OpenAI-compatible endpoint - with no infrastructure to deploy or manage. It is powered by SaladCloud's
distributed GPU network, bringing the same cost-efficiency of self-hosted models with the simplicity of a hosted API.

Any tool or library that supports a custom base URL - Aider, OpenCode, OpenClaw, Cline, Cursor, Goose, Vercel AI SDK,
and more - works with Salad AI Gateway out of the box.

## API Endpoint

```
https://ai.salad.cloud/v1
```

## Available Models

The following models are available during the closed beta:

| Model             | Description                                                                    |
| :---------------- | :----------------------------------------------------------------------------- |
| `qwen3.6-35b-a3b` | Qwen 3.6 35B Mixture of Experts - best for agentic tasks and complex reasoning |
| `qwen3.6-27b`     | Qwen 3.6 27B - strong balance of capability and speed                          |
| `qwen3.5-9b`      | Qwen 3.5 9B - fastest response times, suited for lighter tasks                 |

Additional models will be added as the service moves toward general availability.

## Quick Start

The API follows the OpenAI `v1/chat/completions` format. Once you have access, send requests using your Salad API key in
the `Authorization` header:

```bash theme={null}
curl https://ai.salad.cloud/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <your-salad-api-key>" \
  -d '{
    "model": "qwen3.6-35b-a3b",
    "messages": [
      {"role": "user", "content": "Explain distributed GPU computing in one paragraph."}
    ]
  }'
```

Make sure to replace `<your-salad-api-key>` with your actual Salad API key.

## Use with OpenAI-Compatible Tools

Because Salad AI Gateway exposes a standard OpenAI-compatible endpoint, you can use it with any tool that supports a
custom base URL.

We have integration guides for many popular agentic tools - Cline, Aider, OpenCode, OpenClaw, Kilo Code, Roo Code,
Continue, Goose, Hermes Agent, and more. All of those guides apply here with one difference: **use standard
authentication instead of the `Salad-Api-Key` custom header**. Set your Salad API key as the API key in whichever field
the tool provides (the `Authorization: Bearer` header), and point the base URL to `https://ai.salad.cloud/v1`.

<CardGroup cols={2}>
  <Card title="Cline" icon="code" href="/container-engine/tutorials/agentic-tools/use-cline-with-saladcloud">
    VS Code AI coding agent
  </Card>

  <Card title="Aider" icon="terminal" href="/container-engine/tutorials/agentic-tools/use-aider-with-saladcloud">
    CLI AI pair programmer
  </Card>

  <Card title="OpenCode" icon="terminal" href="/container-engine/tutorials/agentic-tools/use-opencode-with-saladcloud">
    Terminal coding agent
  </Card>

  <Card title="OpenClaw" icon="terminal" href="/container-engine/tutorials/agentic-tools/use-openclaw-with-saladcloud">
    Always-on AI assistant
  </Card>

  <Card title="Kilo Code" icon="code" href="/container-engine/tutorials/agentic-tools/use-kilo-code-with-saladcloud">
    VS Code coding agent
  </Card>

  <Card title="Roo Code" icon="code" href="/container-engine/tutorials/agentic-tools/use-roo-code-with-saladcloud">
    VS Code AI coding agent
  </Card>

  <Card title="Continue" icon="code" href="/container-engine/tutorials/agentic-tools/use-continue-with-saladcloud">
    VS Code + JetBrains coding assistant
  </Card>

  <Card title="Vercel AI SDK" icon="code" href="/container-engine/tutorials/agentic-tools/use-vercel-ai-sdk-with-saladcloud">
    TypeScript AI app framework
  </Card>

  <Card title="Goose" icon="robot" href="/container-engine/tutorials/agentic-tools/use-goose-with-saladcloud">
    CLI + desktop AI agent
  </Card>

  <Card title="Hermes Agent" icon="brain" href="/container-engine/tutorials/agentic-tools/use-hermes-agent-with-saladcloud">
    Self-improving AI agent with MCP
  </Card>
</CardGroup>

**Python (openai SDK):**

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

client = OpenAI(
    base_url="https://ai.salad.cloud/v1",
    api_key="your-salad-api-key",
)

response = client.chat.completions.create(
    model="qwen3.6-35b-a3b",
    messages=[{"role": "user", "content": "Hello from Salad AI Gateway!"}],
)
print(response.choices[0].message.content)
```

**Environment variables (for tools like Aider, etc.):**

```bash theme={null}
export OPENAI_BASE_URL=https://ai.salad.cloud/v1
export OPENAI_API_KEY=your-salad-api-key
```

## Why Salad AI Gateway?

* **No Infrastructure** - no container groups to deploy, no replicas to manage, no cold starts to worry about
* **Always on** - our distributed GPU network ensures high availability and reliability
* **OpenAI-compatible** - easy integration with any OpenAI-compatible tool
* **SaladCloud Pricing** - 90% cost savings vs traditional providers
* **Data Privacy** - your data is never used for training
* **Powered by Sustainable Computing** - utilizing otherwise wasted GPU resources to reduce carbon footprint

## Get Access

Salad AI Gateway is currently in closed beta.

[Sign up for early access](https://salad.com/ai-gateway)