Skip to main content
Zenflow is a desktop app that runs AI agents inside your codebase to implement features, fix bugs, refactor, write tests, and open PRs. It connects to GitHub, Jira, Linear, and your preferred AI coding agent. In this quickstart you’ll:
  • Add a repository to Zenflow
  • Run an agent task with a built-in workflow
  • Review and merge the output
Prerequisite: Ensure you have installed and signed into Zenflow before starting this quickstart.

​
Step 1: Onboarding

On first launch, Zenflow walks you through initial setup:
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 → Saved presets.
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.
Default agent picker showing Zencoder, Claude Code, Codex, and Gemini options
Zenflow confirming whether the selected agent is installed locally
Login guidance for completing the selected agent setup

​
Step 2: Create Your First Task

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.
Task creation screen showing Code/Work toggle, repository selector, description input, agent preset, workflow picker, and Start button
2

Pick a Workflow (2 minutes)

Choose how Zenflow should execute the work:
  • Auto — agent decides the approach (best for most tasks)
  • Fix a Bug — structured investigation → solution → implementation
  • Spec First — technical spec before coding
  • Requirements First — full PRD → spec → staged implementation
  • Multi-model — different models for planning, implementation, review
  • Custom — your own workflow definitions
    Learn how to define custom workflows →
Structured workflows keep results consistent and avoid prompt drift.
3

Verify & Ship (3–6 minutes)

If you configured a verification script, Zenflow runs it after each agent turn (linting, type checks, tests). Review the Changes tab, then merge.
That’s it — you’re using Zenflow.

​
Next Steps

Key Concepts

Understand repositories, tasks, and workflows

Workflows

Explore built-in and custom workflows

Multi-Agent Orchestration

Multi-model pipelines and parallel review

Repository Configuration

Set up verification scripts and presets
https://mintlify.s3.us-west-1.amazonaws.com/forgoodaiinc/quickstart/faq/pricing

Plans & Pricing

Subscription tiers and billing

Supported Agents

Zencoder, Claude Code, Codex, Gemini, and custom agents