Let AI surgically import your code
Open your favourite code editor, connect to Stellify's MCP server, and let AI import your infrastructure—models, migrations, middleware, and more. Then rebuild your controllers and frontend with surgical precision.
Your data layer comes in. Your application logic gets rebuilt properly. No third-party package baggage, no accumulated tech debt—just clean code stored as structured JSON.
Works with all frameworks
What AI imports vs. what you rebuild
AI surgically imports your data layer and infrastructure. Application logic—controllers, services, frontend—gets rebuilt using structured MCP tools for a clean architecture.
AI imports from your codebase
- Models — Eloquent models with relationships, scopes, and casts
- Migrations — Your complete database schema
- Middleware — Request middleware and filters
- Mail, Notifications, Events, Jobs — Laravel core extensions
- Policies, Rules — Authorization and validation
- Config, Blade templates — Configuration and server-rendered views
- Factories, Seeders — Test data generation
AI reads your local files and converts them to Stellify's structured format. Third-party package dependencies are automatically detected and handled.
Rebuilt with MCP tools
- Controllers — AI generates CRUD from your imported models
- API routes — Structured route definitions with middleware
- Form requests — Validation classes based on model fields
- Vue components — Frontend built with structured elements
- Services — Business logic extracted cleanly
AI understands your imported models and generates consistent, well-structured application code. No copy-pasting patterns from your old codebase.
The advantage every text-based tool lacks
Every file, method, statement, and clause in Stellify has a UUID and structured relationships. This means you can ask questions about your codebase that are impossible in a text editor without manually auditing every file.
More importantly, AI can make changes that would be dangerous with text manipulation. Rename a model across 50 files? Text-based AI rewrites each file and hopes line numbers didn't shift. Stellify updates references by UUID—the operation is structural, not textual.
Query your codebase like a senior developer would
These are the questions an experienced developer asks during code review or when onboarding to a new project. Stellify answers them programmatically.
"Which controllers have methods longer than 50 statements?"
Instantly identifies fat controllers that need refactoring. In VS Code you'd be opening every file and eyeballing it.
"Which models have no relationships defined?"
Finds models that are probably incomplete or being used as dumb data containers when they should have proper Eloquent relationships.
"Which controller methods don't use form requests?"
A code quality audit that would take hours becomes instant. Find every store/update method that validates inline instead of using dedicated form request classes.
"Which routes have no middleware?"
Security audit in seconds. Find unprotected routes that should have auth, throttling, or other middleware applied.
"Show me every method that calls DB:: directly"
Finds raw queries that bypass your model layer. These often indicate missing Eloquent relationships or scopes that should be added.
"Which methods reference a model that no longer exists?"
Dead reference detection. Find orphaned code that references deleted models, saving you from runtime errors in production.
"Which services are only used by one controller?"
Finds candidates for inlining or consolidation. Over-abstraction creates cognitive overhead—sometimes the "service" should just be a method.
"Which migration fields aren't in the model's $fillable?"
Compare your migrations against your models. Find fields you're storing but can't mass-assign, which often indicates incomplete model definitions.
Surgical refactoring, not file regeneration
This is where AI working on structured data beats AI working on text files. Every operation targets specific statements by UUID—surrounding code is untouched by definition.
Extract method
Take lines 45-80 of a controller method and move them into a new private method or service class. AI knows exactly which statements to move because they're individually addressable. It creates the new method, moves the statements, adds the call in the original method, and handles variable dependencies.
In a text-based tool: AI rewrites the entire file and hopes the line numbers didn't shift. You review 200 lines of diff to verify 20 lines actually changed.
Move to service
Take business logic out of a controller and into a dedicated service class. AI creates the service file, creates the method, moves the relevant statements, adds the constructor injection in the controller, and updates the method to call the service.
Multiple files, precise changes: Each modification is targeted. The controller's other methods aren't regenerated—only the injection and the calling code change.
Add validation
AI examines a controller's store/update method, analyses the statements to understand what data is being used, and generates a form request class with the right rules. It then updates the controller method signature to type-hint the form request instead of using Request.
Structural understanding: AI doesn't guess what fields you're validating—it reads the statements that use request data and generates rules that match actual usage.
Standardise return types
Find every controller method that returns inconsistent response formats—some return views, some return JSON, some return redirects with no pattern. AI can standardise them to follow your conventions.
Pattern enforcement: Once you've defined the pattern, AI applies it across every method that needs updating. Consistency becomes automatic.
Add missing relationships
AI compares your migration files (which define foreign keys) against your model files (which define Eloquent relationships). If a migration has a user_id foreign key but the model doesn't have a belongsTo relationship, AI adds it.
Cross-file analysis: The relationship between migrations and models is structural. AI doesn't need to parse text—it queries the data.
Cross-file changes with confidence
Because Stellify stores everything as structured data with references, AI can make changes that span multiple files without the risks of text-based manipulation.
Rename a model
Rename Post to Article and have it update the model file, every controller that references it, every route, every migration, every factory, every seeder, and every relationship on other models.
In a text editor this is find-and-replace and prayer. In Stellify the references are structural.
Change a method signature
Add a parameter to a service method and have every caller updated automatically, because Stellify knows which statements in which methods call this method.
No grep, no manual verification, no missed call sites discovered in production.
Delete a route
Remove a route and identify the controller method that's now orphaned, the middleware that might no longer be needed, the form request that's no longer used.
Cleanup becomes comprehensive instead of best-effort.
Maintenance and ongoing improvement
Beyond one-off refactors, there's a maintenance story. These aren't occasional tools—they're part of your regular workflow.
Quality analysis
Check for fat controllers, missing validation, raw DB queries, unused methods, inconsistent naming, and missing return types. Get a prioritised list of improvements, not a wall of warnings.
Performance analysis
Identify N+1 query risks (methods that load a model then access a relationship in a loop without eager loading), oversized queries, and missing indexes based on common query patterns.
Generate missing tests
AI examines controller methods, understands inputs and outputs structurally, and generates feature tests covering happy paths and validation failures. It knows your form request rules, so tests cover actual validation logic.
Consistency checks
When someone adds a new model, check whether it follows established patterns: does it have a factory? A seeder? Relationships matching the migration? A policy if other models have policies?
Frequently asked questions
Common questions about importing and working with existing Laravel projects.
What parts of my project can AI import?
AI surgically imports your entire codebase: models, migrations, controllers, middleware, mail classes, notifications, events, jobs, policies, rules, config files, and templates. Everything gets converted to Stellify's structured format for precise AI operations.
Which code editors work with Stellify?
Any editor that supports MCP (Model Context Protocol) works with Stellify. This includes Claude Code, Cursor, Windsurf, and VS Code with MCP extensions. Connect to our MCP server and AI gets direct access to Stellify's structured tools for surgical code operations.
Can I export my code back to standard files?
Yes. Stellify exports to standard PHP files and Vue components. The exported code is identical to what you'd write by hand—no lock-in, no proprietary syntax. You can work in Stellify for refactoring, then export and deploy normally.
How does AI refactoring compare to doing it manually?
Manual refactoring requires you to track every reference across files, update tests, and hope you didn't miss anything. Stellify's AI operates on structured data—it knows every reference by UUID. When you rename a model, every relationship, migration, and factory updates automatically. The diff shows only targeted changes.
What's the workflow for an existing project?
Open your project in your favourite code editor, connect to Stellify's MCP server, and let AI import your codebase. Then use the MCP tools to refactor, extend, or rebuild any part of your application. AI understands your imported code and generates new code that follows consistent patterns.
Open your editor, connect, and import
Load your favourite code editor, connect to Stellify's MCP server, and let AI surgically import your codebase. Your data layer comes in clean, ready to rebuild properly.
No credit card required