Fragments by E2B Documentation

To run Fragments locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/e2b-dev/fragments.git
   cd fragments

2. Install dependencies:

   npm i

3. Configure environment variables: Create a .env.local file in the root directory and provide your API keys (see Environment Variables Reference).
4. Run the development server:

   npm run dev

   Or build the application:

   npm run build

To add new models or providers to the Fragments UI, modify lib/models.ts.

Adding a Model

Add an entry to the models list in lib/models.ts:

{
  "id": "mistral-large",
  "name": "Mistral Large",
  "provider": "Ollama",
  "providerId": "ollama"
}

Adding a Provider

Add a new configuration to the providerConfigs list in lib/models.ts. For example, to add Fireworks:

fireworks: () => createOpenAI({ apiKey: apiKey || process.env.FIREWORKS_API_KEY, baseURL: baseURL || 'https://api.fireworks.ai/inference/v1' })(modelNameString),

You can also customize the default structured output mode by adjusting the getDefaultMode function for specific providerIds.

This template provides a minimal entrypoint for a Streamlit application. It uses the streamlit library to render a simple 'Hello, World!' message to the UI.

# Scaffolding for a Streamlit app

import streamlit as st

st.write("Hello, World!")

To build a simple Gradio interface, define a function that processes inputs and returns an output, then wrap it in a gr.Interface. You can specify input and output types using strings like "text" or "slider". Finally, call .launch() to start the application.

import gradio as gr

def greet(name, intensity):
    return "Hello, " + name + "!" * int(intensity)

demo = gr.Interface(
    fn=greet,
    inputs=["text", "slider"],
    outputs=["text"],
)

demo.launch()

The fragmentSchema defines the structure for fragment generation data. Use this schema to validate objects containing metadata, code, and dependency information for a fragment.

Key fields include:

• commentary: Detailed description of the generation steps.
• template: The name of the template used.
• title: A short title (max 3 words).
• description: A short description (max 1 sentence).
• additional_dependencies: An array of strings for dependencies not included in the template.
• has_additional_dependencies: Boolean indicating if extra dependencies are needed.
• install_dependencies_command: The command used to install additional dependencies.
• port: The port number used by the fragment (nullable if no ports are exposed).
• file_path: Relative path to the file.
• code: The generated runnable code.

Use createRateLimitResponse to generate a 429 Too Many Requests response that includes standard rate-limiting headers. This is useful for informing clients about their current usage status.

Parameters:

• limit: An object containing:
  • amount (number): The total limit.
  • remaining (number): The number of requests remaining.
  • reset (number): The timestamp or value indicating when the limit resets.