Steamship

Steamship is the fastest way to add Language AI to your software.

Think of us as a package manager for AI. Each Steamship package runs in the cloud on a managed stack. The best way to start is to try one out.

First, install our CLI with the following terminal command (you’ll need NPM):

npm install -g @steamship/cli

Next, select a Steamship package to try with:

ship try

We’ll help you pick a Jupyter Notebook demo you can run on your own machine.

Next Steps

Using Packages

Use full-stack language AI packages in your own code.

Developing Packages

Build and deploy packages with our low-code framework.

Package Cookbook

Package examples for common scenarios.

Contents

• Configuration
  • Authentication
    • Steamship Configuration File
    • Using Multiple Profiles
    • Environment Variables
  • Client Libraries
    • Python Client
    • Typescript Client
  • CLI
  • HTTP API
    • Requests
      • Optional Headers
      • Engine Response Format
    • Creating a Package Instance
    • Invoking a Package Method
• Packages
  • Packages Table of Contents
    • Using Packages
      • Package FAQ
    • Developing Packages
      • Customizing your Package
    • Package Project Structure
    • Package Cookbook
      • Working with Language
      • Building Packages
• Plugins
  • Importers
    • Using File Importers
      • Using a File Importer from within a Steamship Package
      • Using a File Importer from within a Steamship Workspace
      • Using a File Importer as a one-off operation
  • Blockifiers
    • Using Blockifiers
      • Common Blockifiers
      • Using a Blockifier from within a Steamship Package
      • Using a Blockifier from within a Steamship Workspace
      • Using a Blockifier as a one-off operation
    • Developing Blockifiers
      • The Blockifier Contract
      • How to Structure Blocks and Tags
      • Synchronous Example: A Pseudo-Markdown Blockifier
      • Asynchronous Blockifiers
  • Taggers
    • Using Taggers
      • Using a Tagger from within a Steamship Package
      • Using a Tagger from within a Steamship Workspace
      • Using a Tagger as a one-off operation
  • Embedders
    • Using Embedders
  • Additional Links
    • Developing Plugins
      • Plugin Project Structure
      • Developing Importers
      • Developing Blockifiers
      • Developing Taggers
      • Developing Embedders
      • Writing Async Plugins
• Workspaces
  • Creating Workspaces
  • Importing Data
    • Import content directly
    • Import content via a Plugin
  • Blockifying Data
  • Workspace Data Model
    • Files
    • Blocks
    • Tags
      • Ways to use Tags
      • Tag Schemas
      • Block and File Tags
  • Querying Data
    • Usage
    • Language Description
      • Unary Predicates
      • Binary Predicates
      • Binary Relations
      • Conjunctions
      • Special Predicates
• Developer Reference
  • Cloning a Starter Project
  • Steamship Manifest Files
  • Plugin Configuration
  • Python Environment Setup
  • Accepting Configuration
    • Declaring your configuration template
    • Defining and using configuration in your code
  • Storing Secrets
  • Writing Tests
    • Logging
    • Throwing Errors
    • Manual Testing
    • Automated testing
      • Automated testing setup
      • Modifying or removing automated testing
  • Deploying
    • Deploying with the Steamship CLI
    • Deploying via GitHub Actions
      • Automated Deployment Setup
      • Modifying or disabling automated deployments
    • Troubleshooting Deployments
      • The deployment fails because the version already exists
      • The deployment fails because the tag does not match the manifest file
      • The deployment fails with an authentication error
• Python Client Reference
  • steamship package
    • Subpackages
      • steamship.base package
      • steamship.client package
      • steamship.data package
      • steamship.invocable package
      • steamship.plugin package
      • steamship.utils package
    • Module contents
• License
• Authors