Quickstart and Core Concepts

Go from idea to verified, production-ready code within minutes.

Welcome to Zenflow

Zenflow transforms how you build software—from initial concept to production-ready code. This quickstart takes 5-10 minutes and shows you how to: \

• Create a project and define your codebase scope \
• Run AI agents to generate verified code \
• Use workflows with built-in testing and validation \
• Ship production-ready features faster

Install the desktop app

1. Download Zenflow

   Visit the Zenflow page on the Zencoder site and click Download.

2. Run the installer

   Install the desktop app for your operating system using the guided installer.

3. Launch Zenflow

   Open the app once installation finishes to initialize the workspace.

4. Sign in or create an account

   Use your existing Zencoder credentials or create a new account to start orchestrating tasks.

Core Concepts

Zenflow organizes work into a hierarchy so every automation layer stays aligned. Explore each level below.

Projects

Projects

Define scope, repositories, and automation presets. Projects keep Kanban boards, task lists, and settings aligned so downstream work stays consistent. You can run multiple repositories at the same time—each repo becomes its own project with tuned workflows. Even within a single repo, carve out focused projects per area or service so automation stays isolated across workstreams.

Tasks

Tasks

Each task runs inside its own Git worktree with workflow-specific artifacts, dedicated branches, and verification gates for parallel, conflict-free execution.

Subtasks & Chats

Subtasks & Chats

Live, step-aware conversations where agents plan, build, and verify while telemetry shows every command and result so you can intervene or approve quickly.

Onboarding

The first launch of Zenflow includes a short onboarding flow so you start with the right automation partner.

1. Pick a default agent

   Choose between Zencoder, Claude Code, Codex, or Gemini as your default co-pilot. You can switch this preference anytime in Settings → Agents.

2. Zenflow validates your setup

   After you select an agent, Zenflow checks whether its CLI or desktop integration is already installed locally. If something is missing, you get guided installation instructions so the agent can run reliably.

3. Sign in to finalize

   Zenflow then walks you through logging into the chosen agent so tasks can run end-to-end without interruption. Once authenticated, you land in Zenflow ready to create your first workflow.

First Steps (5–10 minutes)

1. Create a Task (2 minutes)

   Start with a concise, outcome-based task description. Examples:

   "Add OAuth for Google and GitHub."
   "Fix checkout bug."
   "Refactor the billing module to reduce duplication."
   

   Focus on what needs to happen, not how to code it.

   
2. Pick a Workflow (2 minutes)

   Choose how Zenflow should execute the work:

   • Quick Change / Fix Bug / Spec and Build / Full SDD Workflow depending on scope

   • Custom workflows tailored to your team’s patterns

     Learn how to define custom workflows →

   Structured workflows keep results consistent and avoid prompt drift.

3. Verify & Ship (3–6 minutes)

   Zenflow’s quality gate ensures only passing code reaches you. Before you review changes, Zenflow can run:

   • Tests
   • Builds
   • Security checks (as configured)

   Multi-agent verification means agents check each other’s work to catch issues early and reduce quality decay.

How Zenflow Stands Apart

Workflows, Not Prompts

Instead of nudging a chatbot repeatedly, Zenflow runs structured workflows that can operate on autopilot or pause for human review checkpoints.

Built-In Verification

Agents validate output before it’s surfaced, so you spend time reviewing good candidates instead of debugging AI output.

Spec-Driven Development

Agents stay anchored to an evolving spec that governs scope, requirements, and quality standards—especially when work spans multiple tasks.

Power Highlights

Parallel Project Execution

Start the next task while agents finish the current one, keeping throughput high even with small teams.

AI-First Project Management

Projects, tasks, subtasks, Kanban views, and inboxes make AI work visible and auditable.

Auto-Generated Task Flows

Workflows break into sequential tasks automatically—autopilot or review-driven—so every run follows a vetted plan.

Model Diversity

Different model families cross-check assumptions to reduce blind spots and improve reliability.