Repository Tags
Customize AI reviews with tags that provide context about your frameworks, languages, and coding standards.
What Are Repository Tags?
Repository tags help Jasper understand the context of your codebase. By tagging repositories with frameworks, languages, and coding styles, you get more relevant and accurate code reviews.
For example, tagging a repository with "Laravel" tells Jasper to:
- Recognize Laravel-specific patterns (Eloquent, Blade, etc.)
- Apply Laravel best practices in reviews
- Understand framework conventions
- Suggest Laravel-idiomatic improvements
Tag Categories
Tags are organized into four categories:
Framework
Web and application frameworks
Language
Programming languages
Tool
Build tools and utilities
Code Style
Coding standards and conventions
Applying Tags to Repositories
To tag a repository:
- Navigate to Settings → Repositories
- Find the repository you want to configure
- Click Edit or the settings icon
- In the Tags section, select applicable tags
- Click Save Changes
You can apply multiple tags from different categories. For example, a repository might have:
- Framework: Laravel, Vue
- Language: PHP, TypeScript
- Tool: Vite, PHPUnit
- Style: PSR-12
System Tags vs Custom Tags
System Tags
Jasper provides a library of pre-configured system tags for popular frameworks, languages, and tools. These tags come with carefully crafted prompts that help the AI understand each technology's best practices and conventions.
Custom Tags
Organizations can create custom tags for internal standards, proprietary frameworks, or specific coding conventions unique to your team.
To create a custom tag:
- Go to Settings → Organization → Tags
- Click Create Custom Tag
- Fill in the tag details:
- Name - Display name (e.g., "Our API Style")
- Category - Framework, Language, Tool, or Style
- Color - Visual identifier
- Prompt Modifier - Instructions for the AI
- Save the tag
Understanding Prompt Modifiers
Each tag has a "prompt modifier" - additional instructions given to the AI when reviewing code in tagged repositories. This is how tags customize the review behavior.
Example: Laravel Tag Prompt
This is a Laravel application. When reviewing:
- Encourage use of Eloquent relationships over raw queries
- Check for proper use of Laravel's validation
- Suggest using Laravel's built-in helpers when appropriate
- Verify proper use of service providers and dependency injection
- Look for N+1 query issues in Eloquent usage
Writing Custom Prompt Modifiers
When creating custom tags, write clear, specific instructions:
- Be specific - "Use camelCase for variables" not "follow naming conventions"
- Give examples - Show what good and bad code looks like
- Prioritize - List most important rules first
- Stay concise - Focus on what's unique to your standard
Pro Tip
Start with system tags and only create custom tags when you need specific behavior not covered by existing tags.
Organization Overrides
Organizations can override the prompt modifier for system tags without creating duplicate custom tags. This is useful when you want to:
- Add organization-specific rules to a framework tag
- Adjust the strictness of style guidelines
- Include internal conventions alongside standard practices
To create an override:
- Go to Settings → Organization → Tags
- Find the system tag you want to modify
- Click Create Override
- Edit the prompt modifier for your organization
- Save the override
Your override only affects repositories in your organization - it doesn't change the system tag for others.
Tag Colors
Tags support a variety of colors for easy visual identification in the UI. Choose colors that help you quickly recognize different types of tags:
Best Practices
Start Simple
Begin with just the primary framework and language tags. Add more specific tags as you learn what feedback is most valuable for your team.
Be Consistent
Apply the same tags to repositories with similar tech stacks. This ensures consistent review feedback across your organization.
Review Tag Effectiveness
Periodically review the feedback Jasper provides. If reviews miss important patterns, consider adding tags or adjusting prompt modifiers.
Don't Over-Tag
More tags aren't always better. Too many conflicting instructions can confuse the AI. Focus on the most important technologies and standards.