Sync Your Codebase

Get your codebase into Scope via GitHub or local file sync.

Two ways to sync

Scope supports two sync methods. Both produce the same structured context.

Option A: GitHub connection

Open your project and go to Settings
Under GitHub Connection, click Connect Repository
Authorize Scope in the OAuth flow
Select the repository you want to analyze
Scope begins analyzing — usually takes 1-2 minutes

Option B: Local sync via MCP

If you prefer not to connect GitHub, or you're working with a local codebase, use the scope_sync MCP tool to send file contents directly:

scope_sync(files: {
  "src/models/user.ts": "export class User { ... }",
  "src/routes/users.ts": "router.get('/users', ...)",
  "prisma/schema.prisma": "model User { ... }"
})

Scope processes the files in memory, extracts structured metadata, and discards the source code. Only changed files are reprocessed on subsequent syncs.

What gets extracted

During analysis, Scope builds a model of your codebase by extracting:

Category	Examples
Data models	User, Order, Product — what data your app stores and manages
Properties	name, price, created date — the details of each data model
Relationships	Orders belong to Users, Products have Reviews
Features	User management, order processing, product catalog — with access rules
Technology	Framework, language, database, test tools, package manager
Patterns	Naming conventions, file organization, coding style

What is NOT stored

Your source code is never stored.

Scope reads files during analysis to extract structure and metadata, but the actual source code is processed in memory and discarded. Only the extracted context (data model names, properties, feature paths, relationships) is stored as searchable context for your project.

Supported languages

Scope can analyze projects in:

Ruby — Rails, schema.rb, ActiveRecord
Python — Django, FastAPI, SQLAlchemy
TypeScript / JavaScript — Express, Next.js, Prisma
Go — Chi, Gin, GORM
Rust — Axum, Actix, Diesel

Scope works with all major database tools and frameworks in these languages. If your stack is listed above, you're covered.

File limits

Maximum file size: 100 KB per file
Maximum files analyzed: 500 files
Files are prioritized by importance (schema, routes, models first)

Keeping context current

Scope supports incremental updates so your context stays current as your codebase evolves:

GitHub — Re-run analysis from project settings anytime
Local sync — Call scope_sync again with updated files. Only changed files are reprocessed.
Via complete_ticket — Include file_contents when completing tickets and Scope re-indexes in the background