Coding with AI
AI assisted coding can boost developer productivity. To ensure the best possible experience, Alokai projects ship with AI contexts. These help AI agents understand the project and write code that is consistent with Alokai best practices.
AI Contexts
AI contexts are structured documentation files that serve as permanent context for AI coding assistants like Cursor, Windsurf, and GitHub Copilot. These files help AI agents:
- Understand the project architecture - How Alokai's middleware, frontend, and integrations work together
- Follow consistent coding patterns - TypeScript conventions, naming patterns, and code structure
- Use proper workflows - How to add new endpoints, extend normalizers, and implement features
- Apply framework-specific knowledge - Alokai's Unified Data Model, SDK usage, and integration patterns
The main context files are:
.cursorrules- context file for Cursor IDE.windsurfrules- context file for Windsurf IDE.github/copilot-instructions.md- context file for GitHub Copilot (VS Code)
They all reference files in ai/ directory which contains information on specific topics.
These files are open for your modifications. If you see that AI constantly gets something wrong, or you introduce new guideline, update the context file accordingly.
Recommended tools
IDEs
GitHub Copilot (VS Code) is supported but not recommended, because it tends to deliver lower quality results.
LLM models
The ai context have been tested with Claude 4.5-sonnet model. However, it doesn't mean it won't deliver good results with other models as well.
Usage
AI context will be automatically loaded by the recommended IDEs. You don't have to do anything special to make it work.. Your IDE will be aware of these contexts and will attach it automatically to your conversations with built-in LLMs.
Prompt Engineering
There are no specific prompt engineering techniques related to Alokai. Follow general prompt engineering best practices. We recommend reading Antrhopic's Prompt Engineering Guide.