Steamship is an SDK and hosting platform for AI Agents and Tools.
Follow our Agent Guidebook for a complete tutorial.
- Build Agents which run in the cloud.
- Use Plugins for common operations like generating text with GPT, converting a CSV to text, or generating an image from text. Steamship manages asynchronicity and retries.
- Store data in Files, Blocks, and Tags. This allows you to query or search it later.
- Deploy as a Package, creating a scalable API for your front end.
- Create as many instances of the Package as you want, each with its own data.
The best way to start is to make a simple package:
Clone our starter repository:
git clone https://github.com/steamship-core/multimodal-agent-starter
Create a virtual environment and install dependencies:
python3 -m venv venv source venv/bin/activate pip install -r requirements.txt pip install -r requirements.dev.txt
ship run local
An ngrok account is required to run this command.
Now that you’ve interacted with your new agent, you’re ready to start modifying it in
See the Agent Guidebook for details.
- Querying Data
- Cloning a Starter Project
- The Steamship Manifest file
- Python Environment Setup
- Accepting Configuration
- Storing Secrets
- Running on Localhost
- Writing Tests
- Deploying with the Steamship CLI
- Deploying via GitHub Actions
- Troubleshooting Deployments
- Monitoring your Instances