Skip to content

steamship-packages/prompt-api-template

Repository files navigation

Overview

This is a basic Replit template for building a Steamship package for prompt-based text generation.

Quick Start

To get started, Click Use Template from the template page in Replit.

To run it:

  1. Get a Steamship API Key (Visit: https://steamship.com/account/api). If you do not already have a Steamship account, you will need to create one.
  2. Copy this key to a Replit Secret named STEAMSHIP_API_KEY.
  3. Click the green Run button at the top of the window (or open a new Shell and type python3 api.py).

Prompt-based Text Generation

This template provides a simple mechanism to generate text based on a supplied prompt. The prompt can ask for any kind of text to be generated, including poems, letters, Twitter responses, LinkedIn posts, jokes, etc. The prompt can be customized based on user input, allowing your users to generate personalized text.

Running the generator

Prerequisites

You will need a Steamship Account and an associated API Key. With an account, an API Key can be obtained via the website (https://www.steamship.com/account/api).

You will need to create a Replit Secret for STEAMSHIP_API_KEY with your Steamship API key.

Local test/development

To run the package locally, use the provided Run button in Replit (or open a new Shell and type python3 api.py).

Production deployment

Once you are satisfied with your prompt, you may deploy it to Steamship in order to create running instance(s) and a demo application.

First, ensure that you have descriptive package handle for your package in steamship.json (defaults to my-prompt-package). This handle will be used by Steamship to refer to your package internally and externally. It must be globally-unique, so you may need to try a few before finding one that is unclaimed.

To deploy it on Steamship, open a new Shell and type ship deploy. This will push a new version of your Steamship package to Steamship's servers.

Details about the package will be accessible via https://www.steamship.com/packages/<your-package-name>. A default demo web app will also be deployed and made globally-accessible.

Package Files

api.py

The main logic for the generator lives entirely in api.py. This file contains a PromptPackage class with one API method: generate(). Inside of generate(), the prompt is built with the user-supplied parameters and submitted to the generation plugin.

steamship.json

After you deploy once, you will see a new steamship.json file in your repl. This is the manifest Steamship uses to store metadata about your package. This includes a handle for the package, as well as a version (allowing for package updates). This information will be used when you deploy your package to Steamship to present package details to users.

The tag Prompt API signals Steamship to automatically generate a special web interface for your package upon instance creation (post-deployment).

requirements.txt

This file is used by Steamship in the package deployment process. Any dependencies that are needed by your code should be added here to ensure they are accessible to your package at runtime.

About Steamship

Steamship is the fastest way to deploy language AI.

Steamship packages provide a mechanism for encapsulating and deploying your language AI logic, allowing invocation with parameters that supports auto-scaling and persistence via a static HTTP endpoint.

Steamship packages can combine multiple Steamship plugins to build complex applications out of various AI building blocks. These plugins cover a broad variety of AI skills, including transcription, summarization, and generation.

To learn more about Steamship packages and plugins, please read our docs: https://docs.steamship.com/.

Feedback and Support

Have any feedback on this template? Or on Steamship in general?

We'd love to hear from you. Please reach out to: [email protected]

About

Template for creating a Prompt API with Steamship

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published