Overview of Vibecraft
Vibecraft is a service that lets you create and publish websites and web applications by communicating with AI in natural language, without programming.
Describe what you want to create in the chat, and Vibecraft will generate the code for a full-fledged application, including frontend and backend, editable in SourceCraft, show a preview, and then publish it to Yandex Cloud.
The service shortens the path from an idea to an MVP or even a full-fledged product: you do not need any special tools or designer/developer skills. For example, with Vibecraft, you can create and publish a landing page, an online course website, a finance tracking app, a sports activity and health tracker, a game, or even a CRM system.
Warning
The service functionality is at the Preview stage and is not covered by the SLA.
Fees are charged for using SourceCraft and Yandex Cloud resources.
Access to the service is provided upon request.
Technical features
The service is automated and uses the following SourceCraft and Yandex Cloud components:
- An ensemble of Code Assistant large language models for developing the application concept, writing code, and testing it.
- A SourceCraft repository for storing code.
- SourceCraft CI/CD for previewing the application and automating deployment.
- A Yandex Container Registry registry for storing Docker images with the application.
- A Yandex Serverless Containers container for running the application.
- A Yandex Managed Service for YDB serverless database for storing the application state.
- A Yandex Lockbox secret for storing sensitive data.
Yandex Cloud resources are managed from the SourceCraft repository using a secure service connection.
Security
Since all projects are hosted in SourceCraft, you can use the following security mechanisms for them:
- Secret scanning: Checking commits for sensitive data contained in the code, e.g., API keys, tokens, certificates, and other secrets.
- Software composition analysis: Identifying dependencies used in software development and known vulnerabilities they may contain.
- Static application security testing analyzes source code to identify vulnerabilities, errors, and security standard violations without executing the actual code.
- AI-powered vulnerability analysis: AI suggestions for detected vulnerabilities, with risk and criticality assessment, false positive probability analysis, and possible fixes.
- Scan history is a dedicated section presenting the history of repository scans performed to detect secrets and analyze dependencies.
For more information, see Security in SourceCraft.