Skip to main content

Your AI Will Follow Your Patterns in 10 Minutes

No complex configuration. No long setup process. Just 3 simple steps.
Time estimate: 10 minutesPrerequisites:
  • ULPI account (sign up free)
  • Repository with code
  • AI assistant with MCP support (Claude Desktop, Cursor, etc.)

Step 1: Connect Your Repository

1

Navigate to Repositories

DashboardRepositories → Click Connect Repository
Pro tip: Connect your most active repository first
2

Choose Your Git Provider

  • GitHub (most common)
  • GitLab (cloud or self-hosted)
  • Bitbucket (cloud or server)
  • Gitea (self-hosted)
Read-only access: ULPI only needs read permissions to detect your tech stack
3

Authorize ULPI

Grant ULPI access to:
  • ✅ Read repository contents
  • ✅ Detect technologies (package.json, composer.json, etc.)
  • No write access required
We DO NOT:
  • Write to your repositories
  • Execute code
  • Store your source code
4

Tech Stack Auto-Detection

ULPI automatically detects your technologies:Scans:
  • package.json → Node.js, React, Next.js
  • composer.json → PHP, Laravel
  • requirements.txt → Python, Django, FastAPI
  • pom.xml → Java, Spring Boot
  • File extensions → Languages, frameworks
Result:
⏱️ Time: 2 minutes

Step 2: Configure Your AI Assistant

Add ULPI Skills MCP server:
  1. Open Claude Desktop settings
  2. Navigate to DeveloperMCP Servers
  3. Click Add Server
  4. Use these settings:
  1. Get your API key:
    • Dashboard → SettingsAPI KeysCreate Key
    • Select scope: Skills
  2. Get your Project ID:
    • Dashboard → Select project → URL shows /projects/123 (123 is your ID)
  3. Restart Claude Desktop
Verify: Start a new chat. Ask: “What skills are available?”Claude should list your matched skills.
⏱️ Time: 5 minutes

Step 3: Use Your First Skill

Ask your AI:
AI response:
⏱️ Time: 2 minutes

✅ You’re Done! What’s Next?

Try More Skills

Explore all 8 skills matched to your project

Create Custom Skill

Encode your team’s patterns:
  • Your deployment workflow
  • Company coding standards
  • Code review checklist
  • Security scanning process
Learn How to Create Custom Skills →

View in Dashboard

Dashboard → Your Project → Skills
  • See all available skills
  • View skill details
  • Check usage statistics
  • Manage active skills

Integrate with Workflow

Combine skills for complete workflows:
See Common Workflows →

Common Workflows

Complete API feature in 10 minutes:Step 1: Create endpoint
→ Generates controller, validation, resource, service, routes, testsStep 2: Add tests
→ Adds authorization tests, validation tests, error handlingStep 3: Document API
→ Generates OpenAPI spec, example requests/responsesStep 4: Commit
→ Creates proper commit message following Conventional CommitsResult:
  • Complete API endpoint
  • 100% test coverage
  • Full documentation
  • Professional commit message
  • Time: 10 minutes (vs. 2 hours manually)
Add new database table:Step 1: Create migration
→ Generates migration with proper types, indexes, foreign keysStep 2: Create model and relationships
→ Generates model with fillable, casts, relationshipsStep 3: Add factory and seeder
→ Generates realistic test dataStep 4: Test migration
→ Creates migration tests, relationship testsResult:
  • Production-ready schema
  • Proper indexes for performance
  • Test data generation
  • Migration tests
Build new UI component:Step 1: Create component
→ Generates component, TypeScript types, CSS module, tests, storyStep 2: Add to Storybook (if you have custom skill)
→ Creates Storybook stories with all variantsStep 3: Test accessibility
→ Adds ARIA label tests, keyboard navigation testsStep 4: Commit
Result:
  • Production-ready component
  • Full TypeScript types
  • Component tests
  • Storybook documentation
  • Accessibility compliant
Deploy to staging (custom skill example):Prerequisites: Create custom “Deploy to Staging” skillUsage:
Skill executes:
Result:
  • Zero mistakes
  • Every step verified
  • Team notified
  • Time: 3 minutes (vs. 20 minutes manually)
Before creating PR (custom skill):Prerequisites: Create custom “Pre-Review Checklist” skillUsage:
Skill checks:
Result:
  • 78% reduction in code review iterations
  • Faster PR approvals
  • Higher code quality

Troubleshooting

Problem: AI says “No skills available”Diagnosis:
  1. Check MCP connection:
    • Restart your AI assistant
    • Start new chat
    • Ask: “What MCP servers are connected?”
  2. Verify API key:
    • Dashboard → Settings → API Keys → Verify key is active
  3. Verify project ID:
    • Dashboard → Select project → Check URL
  4. Check API key scope:
    • API key must have Skills scope enabled
    • Dashboard → Settings → API Keys → Check scopes
Fix:
  • Correct API key or project ID in MCP config
  • Restart AI assistant
  • Test: “What skills are available?”
Problem: Seeing Laravel skills in React project (or vice versa)Cause: Tech stack not detected correctlyFix:
  1. Check detected technologies:
    • Dashboard → Your Project → Tech Stack
  2. Manually add missing technologies:
    • Click Edit Tech Stack
    • Add: React, TypeScript, Node.js, etc.
    • Save
  3. Reconnect repository:
    • Sometimes helps re-trigger detection
    • Dashboard → Repositories → Disconnect → Reconnect
  4. Verify package files exist:
    • React projects need package.json with react dependency
    • Laravel projects need composer.json with laravel/framework
Result: Skills filtered correctly for your stack
Problem: AI says “Skill execution failed”Common causes:1. Missing dependencies:
→ Skill requires Laravel, but project is not Laravel2. Invalid parameters:
→ Check skill requirements (e.g., endpoint must start with /)3. Quota exceeded:
→ Upgrade plan or wait for monthly reset4. File conflicts:
→ Skill tries to create existing fileFix:
  • Read error message carefully
  • Check skill prerequisites
  • Verify project structure
  • Ask AI to retry with corrections
In Dashboard:
  1. Dashboard → Your Project → Skills
  2. Click on skill name
  3. View:
    • Full skill description
    • Step-by-step guide
    • Prerequisites
    • Example usage
    • Related skills
In AI Assistant:
AI will show:
  • What the skill does
  • What files it creates
  • Prerequisites
  • Example usage
View your usage:
  1. Dashboard → BillingUsage
Shows:
  • Skill runs this month
  • Runs by skill (which skills used most)
  • Tokens consumed
  • Remaining quota
Example:
Tip: Upgrade to Pro if consistently exceeding limits

Quick Reference


What’s Next?

1

Explore All Skills

Ask your AI: “Show me all available skills with examples”
2

Learn How Skills Work

Technical deep dive into skill execution and token usage
3

Create Custom Skills

Encode your team’s patterns into reusable workflows
4

Optimize Your Workflow

Combine skills for end-to-end workflows (API → Tests → Docs → Commit)

Need Help?

Documentation

Technical guides

Support

Email us anytime

Community

Join our Discord