Skip to content

Development

Guidelines for developers working on the platform, organized by intent: get productive, follow our standards, or learn the theory.


Getting Started

New to the project? Start here.

  • Environment Setup — Local environment, dependencies, and tooling
  • Workflows — Git branching, code review, CI/CD, and release process

Tooling & Processes

Tools and workflows that support daily development.

Section Description
Issue Tracking YouTrack projects, fields, workflows, and writing good issues
AI-Assisted Development Claude Code setup, daily workflows, skills, and agent teams
Shell & CLI Daily commands, Docker, debugging, scripts, and productivity
CI/CD GitHub Actions, build/test pipelines, deployment, and environments

Engineering Standards

How we write and ship code across the stack.

Section Description
Backend Python / FastAPI architecture, patterns, and reliability
Frontend TypeScript / React / Next.js standards, patterns, and frameworks
Database PostgreSQL migrations, connection management, and operational patterns
API Design OpenAPI, versioning, error contracts, and documentation
Testing Backend and frontend testing strategies
Security Authentication, authorization, secrets, input validation, OWASP
Monitoring Logging, metrics, tracing, and alerting

Cookbook

Step-by-step recipes for common extension points.

Recipe What You'll Build
Add a Generation Strategy New strategy in the generation pipeline
Add a Task Type New compute task from model to dispatch map
Add an AI Provider New provider processor with queue routing
Concurrency Recipes Common concurrent and parallel task patterns

Deep Dives

Theory, internals, and advanced patterns for when you need to go deeper.

Section Description
Concurrency Processes vs threads, Python GIL, asyncio, event loop, worker pools
Database Internals SQL fundamentals, query performance, indexing, transactions, backup
Performance Profiling, load testing, benchmarking, optimization patterns
Advanced Patterns Real-time features, server actions, streaming, event-driven, feature flags

Key Principles

  1. Consistency Over Innovation — Follow existing patterns unless there's a compelling reason to change
  2. Test-Driven Development — Write tests while implementing, not after
  3. Atomic Implementation — Break features into small, testable tasks
  4. Pattern-First — New features follow existing patterns; AI agents are guided by referencing existing implementations
Goal Path
New backend developer Getting Started → Standards/Backend → Deep Dives/Database
New frontend developer Getting Started → Standards/Frontend → Deep Dives/Advanced
Building a new feature Cookbook → Standards (relevant stack) → Deep Dives as needed
Setting up AI tools Tooling/AI-Assisted → Getting Started