API Reference

Welcome to the TigerCity.ai API Reference. Below is an overview of all available endpoints.

How API Endpoints Work

API endpoints are URLs that your application can call to interact with TigerCity.ai services. Each endpoint performs a specific action, such as creating a model, getting a response, or managing your account. You make HTTP requests (GET, POST, DELETE, etc.) to these endpoints with your API token for authentication.

Obtaining an API Token

To use the TigerCity.ai API, you need an API token for authentication. Here's how to get one:

Go to the Playground section
Navigate to ApiKeys
Click the (+) button to create a new API key
Give it a name and you will get a dialog showing the new key
Copy the key and set it in your environment

Important: When the key is generated, it is only shown once and will not be displayed again. Make sure to save it immediately, as you won't be able to retrieve it later. If you lose the key, you'll need to create a new one.

Example:

export TIGER_API_KEY=<your app key>

Setting the API Key Permanently

Linux / macOS

Add the export command to your shell configuration file:

For bash: Add to ~/.bashrc
For zsh: Add to ~/.zshrc

Open the file in a text editor and add:

export TIGER_API_KEY=<your_api_key_here>

Then reload your shell configuration: source ~/.bashrc (or source ~/.zshrc for zsh)

Windows

For Command Prompt:

Open System Properties → Advanced → Environment Variables
Under "User variables", click "New"
Variable name: TIGER_API_KEY
Variable value: your API key
Click OK and restart your terminal

For PowerShell:

Run:

[System.Environment]::SetEnvironmentVariable('TIGER_API_KEY', 'your_api_key_here', 'User')

Restart PowerShell for the changes to take effect

Streaming

Many endpoints support streaming responses for real-time feedback. Learn more about how streaming works and how to implement it in your application: Streaming Guide

Available Endpoints

Generate Response

v1/generate_response

Send a structured list of input messages with text, and the model will generate the next message in the conversation.

List Models

v1/models

List all available models, both the base models (Llama-3.1-8b and DeepSeek-R1-8b) and all the private models created by the current user.

Create Model

v1/models

Create Model allows you to create a copy of an existing model (called the parent model). Once created, your new model is independent and can be customized through training. You can experiment with different training parameters, datasets, and fine-tuning techniques without any risk to the original parent model.

Note: You can only train models that you have created yourself. The standard base models (like Llama-3.1-8b or DeepSeek-R1-8b) are read-only, so you must first create a copy of these models before you can train them.

Delete Model

v1/models/{model-id}

Delete a model that you have created. This permanently removes the model from your account.

Note: You can only delete models that you have created yourself. The standard base models (like Llama-3.1-8b or DeepSeek-R1-8b) are read-only and cannot be deleted.

Train

v1/train

Fine-tune a language model by adjusting its weights to make it more likely to produce the results you want. Training modifies the model's internal parameters based on your provided examples, enabling it to learn new behaviors, preferences, and response patterns.

To train the model, you provide a conversation context (messages) along with multiple possible responses, each with a reward value indicating how desirable that response is. The model learns from these examples by updating its weights to favor responses with higher rewards, gradually improving its ability to generate the types of outputs you prefer.

This process allows you to customize the model's behavior for specific tasks, styles, or preferences without needing to build a model from scratch.