together Plugin for TSDIAPI

A TSDIAPI plugin to extend API functionality with Together AI.

📌 About

This is a TSDIAPI plugin designed to integrate Together AI models into your API. It provides an easy way to interact with Together AI for text generation, image analysis, and more.

🔗 TSDIAPI CLI: @tsdiapi/cli

📦 Installation

Install the plugin using the TSDIAPI CLI:

tsdiapi plugins add together

You can install this plugin using npm:

npm install --save @tsdiapi/together

Then, register the plugin in your TSDIAPI project:

import { createApp } from "@tsdiapi/server";
import createPlugin from "@tsdiapi/together";

createApp({
  plugins: [createPlugin()],
});

🚀 Features

• 🛠 Seamless integration with Together AI inside TSDIAPI.
• 🎨 Supports text and image-based AI models.
• ⚙ Highly configurable with environment variables and plugin options.
• 📊 Tracks token usage for cost management.

🔧 Configuration Options

This plugin supports both direct options and environment variables for flexible configuration.

Plugin Options

createPlugin({
  apiKey: "your-api-key",
  model: "meta-llama/Llama-Vision-Free",
  maxTokens: 200,
  temperature: 0.3,
  topP: 0.7,
  topK: 50,
  repetitionPenalty: 1.2,
});

| Option | Type | Default | ENV Variable | Description | | ------------------- | -------- | -------------------------------- | ----------------------------- | ----------------------------------------------------- | | apiKey | string | "" | TOGETHER_API_KEY | Together AI API key (Required) | | model | string | "meta-llama/Llama-Vision-Free" | TOGETHER_DEFAULT_MODEL | Default Together AI model to use | | maxTokens | number | 200 | TOGETHER_MAX_TOKENS | Maximum number of tokens per response | | temperature | number | 0.3 | TOGETHER_TEMPERATURE | Sampling temperature for response randomness | | topP | number | 0.7 | TOGETHER_TOP_P | Top-p (nucleus sampling) value | | topK | number | 50 | TOGETHER_TOP_K | Top-k sampling value | | repetitionPenalty | number | 1.2 | TOGETHER_REPETITION_PENALTY | Repetition penalty for controlling repeated responses |

Using Environment Variables

Instead of defining options in the code, you can set them as environment variables:

TOGETHER_API_KEY=your-api-key
TOGETHER_DEFAULT_MODEL=meta-llama/Llama-Vision-Free
TOGETHER_MAX_TOKENS=200
TOGETHER_TEMPERATURE=0.3
TOGETHER_TOP_P=0.7
TOGETHER_TOP_K=50
TOGETHER_REPETITION_PENALTY=1.2

📌 How to Use

After installation, you can use the Together AI provider to generate text and analyze images.

Get Together Provider

import { useTogetherProvider } from "@tsdiapi/together";

const together = useTogetherProvider();

Basic Chat Completion

const response = await together.chat("Tell me a joke!");
console.log(response.result);

Analyze an Image

const response = await together.analyzeImage(
  "https://example.com/image.jpg",
  "Describe this image."
);
console.log(response.result);

🔗 Related Plugins

You can find more TSDIAPI plugins here:
🔗 Available Plugins

👨‍💻 Contributing

Contributions are always welcome! Feel free to submit issues or pull requests to improve this plugin.

📧 Contact: [email protected]

🚀 Happy coding with TSDIAPI & Together AI! 🎉