@elizaos/plugin-gateway

Vercel AI Gateway plugin for elizaOS. Provides access to AI models through the Vercel AI Gateway with zero markup.

Features

- Text Generation: TEXT_SMALL, TEXT_LARGE chat completions
- Embeddings: Vector embeddings generation
- Images: Image generation and description
- Structured Output: JSON object generation
- Streaming: Real-time streaming support

Multi-Language Support

This plugin is available in three languages with full feature parity:

| Language | Directory | Package Name |
| ---------- | ------------- | ------------------------- |
| TypeScript | typescript/ | @elizaos/plugin-gateway |
| Python | python/ | elizaos-plugin-gateway |
| Rust | rust/ | elizaos-plugin-gateway |

Installation

$3

``bash
npm install @elizaos/plugin-gateway


or


bun add @elizaos/plugin-gateway
`

$3

`bash
pip install elizaos-plugin-gateway
`

$3

`toml
[dependencies]
elizaos-plugin-gateway = "1.0"
`

Quick Start

$3

`typescript
import { gatewayPlugin } from "@elizaos/plugin-gateway";
import { createAgent } from "@elizaos/core";

const agent = createAgent({
plugins: [gatewayPlugin],
});

// Use via runtime
const text = await runtime.useModel(ModelType.TEXT_LARGE, {
prompt: "What is quantum computing?",
});
`

$3

`python
from elizaos_plugin_gateway import GatewayPlugin

async with GatewayPlugin() as plugin:
response = await plugin.generate_text_large("What is quantum computing?")
print(response)
`

$3

`rust
use elizaos_plugin_gateway::{get_gateway_plugin, TextGenerationParams};
use futures::StreamExt;

let plugin = get_gateway_plugin()?;

// Standard text generation
let response = plugin.generate_text("What is quantum computing?").await?;
println!("{}", response);

// Streaming text generation
let params = TextGenerationParams::new("Tell me a story");
let mut stream = plugin.stream_text(¶ms).await?;
while let Some(chunk) = stream.next().await {
print!("{}", chunk?);
}
`

Configuration

Set environment variables:

`bash


Required - one of:

export AI_GATEWAY_API_KEY="your-api-key"
export AIGATEWAY_API_KEY="your-api-key"
export VERCEL_OIDC_TOKEN="your-oidc-token"

Optional

export AI_GATEWAY_BASE_URL="https://ai-gateway.vercel.sh/v1"
export AI_GATEWAY_SMALL_MODEL="gpt-5-mini"
export AI_GATEWAY_LARGE_MODEL="gpt-5"
export AI_GATEWAY_EMBEDDING_MODEL="text-embedding-3-small"
export AI_GATEWAY_EMBEDDING_DIMENSIONS="1536"
export AI_GATEWAY_IMAGE_MODEL="dall-e-3"
`

Supported Model Types

| Model Type | Description |
| ------------------- | ------------------------------- |
| TEXT_SMALL | Fast, efficient text generation |
| TEXT_LARGE | High-quality text generation |
| TEXT_EMBEDDING | Vector embeddings |
| OBJECT_SMALL | JSON generation (small model) |
| OBJECT_LARGE | JSON generation (large model) |
| IMAGE | Image generation |
| IMAGE_DESCRIPTION | Image analysis |

Development

$3

`bash


TypeScript

bun run build:ts

Python

bun run build:python

Rust

bun run build:rust
`

$3

`bash


TypeScript

bun run test:ts

Python

bun run test:python

Rust

bun run test:rust
`

$3

`bash
bun run lint
`

Architecture

`
plugin-vercel-ai-gateway/
├── typescript/ # TypeScript implementation
│ ├── index.ts # Main plugin export
│ ├── types/ # Type definitions
│ ├── models/ # Model handlers
│ ├── providers/ # HTTP client
│ └── utils/ # Configuration utilities
├── python/ # Python implementation
│ ├── elizaos_plugin_gateway/
│ │ ├── __init__.py
│ │ ├── client.py # HTTP client
│ │ ├── config.py # Configuration
│ │ ├── plugin.py # Plugin class
│ │ └── types.py # Type definitions
│ └── tests/
└── rust/ # Rust implementation
├── src/
│ ├── lib.rs # Main export
│ ├── client.rs # HTTP client
│ ├── config.rs # Configuration
│ ├── error.rs # Error types
│ └── types.rs # Type definitions
└── tests/
`

License

MIT

Contributing

1. Fork the repository
2. Create your feature branch: git checkout -b feature/amazing-feature
3. Commit your changes: git commit -m 'Add amazing feature'
4. Push to the branch: git push origin feature/amazing-feature`
5. Open a Pull Request

Please ensure all three language implementations maintain parity.