@playai/aura

Getting Started

This project is initialized with Bun.

Installation

Linux & macOS

curl -fsSL https://bun.sh/install | bash

Windows

powershell -c "irm bun.sh/install.ps1 | iex"

Setup

1. Install dependencies:

   bun install

2. Setup the database:

   bun run migrate

   This initializes a persistent pglite database with local storage. PgLite is a lightweight, embedded PostgreSQL-compatible database designed for local development and testing.

3. Open database explorer using Drizzle:

   bun run studio

Running the Server

• To start the development server:

  bun run dev

• To start the server:

  bun run start

• You can view the API documentation at the /docs endpoint.

Aura Setup

1. Create an Aura service:

   To create an Aura service, you need to make an API request to the /aura/create endpoint. The request should be a POST request with the following JSON payload:

   • name: The designated name of the service.
   • description: A brief description of the service.
   • headers: HTTP headers that are automatically injected from the backend during the API call.
   • input: Input fields expected from the user and it's type.
   • hiddenInput: Sensitive input fields required for the API call but not visible to the user.
   • endpoint: The API endpoint URL.
   • requestType: The HTTP request method, which can be either POST or GET.
   • payment: An optional value representing the amount of SOL tokens the user must transfer to utilize the service. If not set, no payment is required.

   Example:

   curl http://localhost:3000/aura/create \
   --request POST \
   --header 'Content-Type: application/json' \
   --data '{
   "name": "Rogue Video Generation",
   "description": "Create a video by agent Rogue based on the given topic",
   "headers": {
   "Authorization": "Bearer ${YOUR_BEARER_TOKEN}"
   },
   "input": {
   "message": "string",
   },
   "hiddenInput": {},
   "endpoint": "http://example.com/api",
   "requestType": "POST",
   "payment": 0.05
   }'

Note

The payment will be sent to the associated wallet with the agent to which this service is linked. This is not shown here since it's a minimal implementation.

2. Retrieve all Aura services:

   To retrieve a paginated list of all Aura services, make a GET request to the /aura/services endpoint. You can use the optional page and limit query parameters to control pagination.

   Example:

   curl http://localhost:3000/aura/services?page=1&limit=10

   {
     "total": 1,
     "page": 1,
     "limit": 10,
     "result": [
       {
         "id": "5086985c-9fa5-4eb3-b0d5-347880609c3f",
         "name": "Rogue Video Generation",
         "description": "Create a video by agent Rogue based on the given topic",
         "input": {
           "message": "string"
         },
         "payment": 0.05
       }
     ]
   }

   The frontend will use this to render all the services with the relevant input fields such as string, number, or file.

3. Delete an Aura service:

   To delete an Aura service, make a DELETE request to the /aura/service/{id} endpoint. Replace {id} with the ID of the service you want to delete.

   Example:

   curl 'http://localhost:3000/aura/service/{id}' \
   --request DELETE

4. Call an Aura service:

   To call an Aura service, make a POST request to the /aura/service/{id} endpoint. Replace {id} with the ID of the service you want to call. The request should include the necessary data mentioned in the input field of the service creation.

   Example:

   curl 'http://localhost:3000/aura/service/{id}' \
     --request POST \
     --header 'Content-Type: application/json' \
     --data '{
     "message": "Talk about the rug pull by Javier Milei & Hayden Davis",
     "signature": "4pfgaYWuKLTUNKBmrra4vzXCWvAMCH4K6YrbQEMHos3ArAqoreARgqGWVBzxr5VL65ZiPxeVk7t65fp3n2ZGRi8z"
   }'

   {
     "video": "http://example.com/rogue.mp4"
   }

Note

The signature field is optional and only required if the service requires payment.

Note

Currently, this API returns the raw results from the endpoint to the frontend. To ensure consistent presentation of results, developers should adhere to a standardized response format in future implementations.