Core Concepts

Flows — The Heart of DevFlow

A Flow is the central unit of work in DevFlow. Every feature, bug fix, or task is a Flow. Unlike simple tickets or issues, a Flow has a structured lifecycle that enforces quality at every step.

The Flow Lifecycle

Every Flow moves through seven states in three phases:

Preparation          Development           Complete
┌─────────────────┐  ┌──────────────────┐  ┌──────┐
│ idea → planning │  │ ready →          │  │      │
│      → approval │  │   in_progress →  │  │ done │
│                 │  │   review         │  │      │
└─────────────────┘  └──────────────────┘  └──────┘

State	What happens	Who acts
idea	A new idea is captured — no details yet	Anyone
planning	The AI agent analyzes requirements and creates a detailed implementation plan	AI Agent
approval	You review the plan. Approve it or send it back with feedback	You
ready	Plan is approved, waiting for implementation to start	—
in_progress	The AI agent writes code following the approved plan	AI Agent
review	You test the implementation and review the results	You
done	Work is complete, code is merged	—

The Key Principle: Humans Decide, AI Executes

DevFlow enforces a clear separation of concerns:

You decide:

Whether an idea is worth pursuing
Whether a plan is good enough (approval)
Whether the implementation meets your standards (review)

The AI agent executes:

Creating detailed implementation plans (planning)
Writing code based on approved plans (in_progress)
Following your project’s coding standards and conventions

This means you always stay in control. The AI never makes irreversible decisions without your approval.

Audit Trail

Every state transition is tracked with who and when:

Event	Recorded
Plan created	Who created it, when
Plan approved	Who approved it, when
Implementation started	Which agent, when
Code approved	Who approved it, when

This gives you full traceability — months later, you can look at any Flow and understand exactly what happened.

Projects

Projects group related Flows together. Each project can have:

Its own Kanban board for visual management
Git settings with branch naming and PR conventions
Release management for versioned delivery
Time tracking aggregated across all Flows
Dashboard with project-specific statistics

Organizations

Organizations add team capabilities:

Members with roles (Owner, Admin, Member)
Member types (Employee, Freelancer, External)
Compliance profiles for labor law enforcement
Shared projects visible to all org members

Users can switch between their personal context (own projects, own license) and organization context (shared projects, org rules).

Time Tracking

DevFlow includes built-in time tracking:

Timer — Start/stop with automatic flow assignment
Manual entries — Log time after the fact
Takt system — Round time to configurable intervals (5, 10, 15, 30 minutes)
Agent tracking — AI agent work time is tracked separately