https://github.com/aditya-nadkarni/spongecake
Spongecake is the easiest way to launch computer use agents.
https://github.com/aditya-nadkarni/spongecake
ai-agents ai-agents-framework automation computer-use docker llm openai python
Last synced: about 2 months ago
JSON representation
Spongecake is the easiest way to launch computer use agents.
- Host: GitHub
- URL: https://github.com/aditya-nadkarni/spongecake
- Owner: aditya-nadkarni
- License: mit
- Created: 2025-03-08T00:54:44.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-04-25T22:24:49.000Z (about 1 year ago)
- Last Synced: 2025-09-04T11:00:53.405Z (8 months ago)
- Topics: ai-agents, ai-agents-framework, automation, computer-use, docker, llm, openai, python
- Language: JavaScript
- Homepage:
- Size: 27.7 MB
- Stars: 149
- Watchers: 2
- Forks: 19
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
Open Source Operator for Computer Use
![[coming soon] Shows a demo of spongecake in action](./static/spongecake-demo.gif)
Using spongecake to automate linkedin prospecting (see examples/linkedin_example.py)
[](https://pypi.org/project/spongecake/)
[](https://docs.spongecake.ai/quickstart)
[](https://img.shields.io/github/stars/aditya-nadkarni/spongecake)
[](https://discord.gg/3KDmjAd98w)
## What is spongecake?
🍰 **spongecake** is the easiest way to create your own OpenAI Operator that can use computers. These agents can scrape, fill out forms, and interact with websites or local apps. It helps:
- Developers who want to build their own computer use / operator experience
- Developers who want to automate workflows in desktop applications with poor / no APIs (super common in industries like supply chain and healthcare)
- Developers who want to automate workflows for enterprises with on-prem environments with constraints like VPNs, firewalls, etc (common in healthcare, finance)
---
# Quick Start
1. **Clone this repository** (if you haven’t already):
```bash
git clone https://github.com/aditya-nadkarni/spongecake.git
cd spongecake
```
2. **Run the setup script**:
```bash
chmod +x setup.sh # May be required on Unix/macOS
./setup.sh
```
3. **Run the backend and frontend**:
In the spongecake directory, run:
```bash
cd spongecake-ui
cd frontend
npm run dev
```
In a new terminal, run:
```bash
cd spongecake-ui
cd backend
python server.py
```
> **Note:** This deploys a Docker container in your local Docker environment. If the spongecake default image isn't available, it will pull the image from Docker Hub.
---
# Documentation
See [full documentation](https://docs.spongecake.ai/quickstart) for more details and examples.
# Demos
## LinkedIn Prospecting
![[coming soon] Shows a demo of spongecake in action](./static/linkedin-example.gif)
Using spongecake to automate linkedin prospecting (see examples/linkedin_example.py)
## Amazon Shopping
![[coming soon] Shows a demo of spongecake in action](./static/amazon-example.gif)
Using spongecake to automate amazon shopping (see examples/amazon_example.py)
## Form Filling
![[coming soon] Shows a demo of spongecake in action](./static/data-entry-example.gif)
Using spongecake to automate form filling (see examples/data_entry_example.py)
# Telemetry
By default, spongecake will send anonymous usage data to help us improve the product. No identifying information is collected. You can disable this by setting the `SPONGECAKE_DISABLE_TELEMETRY` environment variable to `true`, or `SPONGECAKE_TELEMETRY` to `false`.
# Appendix
## Contributing
Feel free to open issues for any feature requests or if you encounter any bugs! We love and appreciate contributions of all forms.
### Pull Request Guidelines
1. **Fork the repo** and **create a new branch** from `main`.
2. **Commit changes** with clear and descriptive messages.
3. **Include tests**, if possible. If adding a feature or fixing a bug, please include or update the relevant tests.
4. **Open a Pull Request** with a clear title and description explaining your work.
## Roadmap
- Support for other computer-use agents
- Support for browser-only environments
- Integrating human-in-the-loop
- (and much more...)
## Appreciation
- Shout out to [assistant-ui](https://github.com/assistant-ui/assistant-ui) for making such an easy to use chat component
## Team
Made with 🍰 in San Francisco