LanguageEnglish
Choose a languageApplies to the public website
EnglishFrançaisFrenchDeutschGermanעבריתHebrewРусскийRussian日本語JapaneseالعربيةArabic简体中文Chinese (Simplified)繁體中文Chinese (Traditional)

Use cases

Start with the work.
Follow the real flow.

The landing page explains the idea. This page shows the actual work: 13 narrated flows captured from the real AirCode Ø interface, from the first agent session to an app people can install.

Start with session launchRead the quickstart
  • 13narrated workflows
  • Real UIproduction interaction paths

Choose an outcome

From one session to an operating system for agent work.

Start with the phase that matches your current job. The same host connects all four paths.

01

Start & operate

Launch across repositories, keep parallel work alive, answer from any screen and schedule what repeats.5 real workflows →02

Build & verify

Follow execution, inspect evidence and connect planned work to isolated agent runs.3 real workflows →03

Understand & reuse

Measure where resources go, check each project’s health and turn verified history into reusable agent context.3 real workflows →04

Build & distribute

Turn one sentence into a running app, then publish it for others to try and install.2 real workflows →
Start & operate

Launch across repositories, keep parallel work alive, answer from any screen and schedule what repeats.

01

Start & operate

Launch a real coding session

Choose the project, harness, model, effort and permissions, then send the first prompt from the actual AirCode Ø workbench.
Use this when

A task needs a deliberate harness, model, runtime and permission setup.

Outcome

A reproducible session that can be reopened from any connected AirCode Ø client.

  1. 1

    Select a repository available on the host.

  2. 2

    Set the harness, model, effort and permissions.

  3. 3

    Launch once and keep the complete session state.

02

Start & operate

Manage repositories in parallel

Keep work across a dozen repositories visible in one interface, arrange sessions in resizable groups, and create named docks without restarting anything.
Use this when

You are coordinating several repositories and separate IDE windows no longer scale.

Outcome

One workbench shows every repository and agent while persistent host sessions remain reachable from connected machines.

  1. 1

    Open agents from independent repositories as workbench tabs.

  2. 2

    Drag tabs into groups and resize every split.

  3. 3

    Create named docks and return to their saved client layouts.

03

Start & operate

Monitor every active session

Use one Active Sessions Dashboard to find working, waiting-for-you and ready agents across every repository, then inspect their latest replies.
Use this when

More repositories and sessions are active than you can sensibly monitor one window at a time.

Outcome

One operational view shows every repository’s progress, attention requests and latest agent returns.

  1. 1

    See working, waiting and ready session counts.

  2. 2

    Read each agent’s latest response in place.

  3. 3

    Filter attention requests and reopen the exact session.

04

Start & operate

Answer agents from your phone

Leave your desk while every session keeps running on your machine. When an agent needs a decision, open it on your phone, reply in the same conversation and let the work resume.
Use this when

An agent needs a decision while you are away from your desk.

Outcome

The session continues on the host with your answer, and every connected screen shows the same conversation.

  1. 1

    See every running agent from your phone.

  2. 2

    Open the session that is waiting for you.

  3. 3

    Reply, and the agent picks up where it stopped.

05

Start & operate

Schedule the work that repeats

Save recurring work in Scheduled Tasks as a workflow whose branches run side by side, and let the Prompt Queue send a single instruction at the hour you choose.
Use this when

The same checks should run every night, or the next instruction should leave later.

Outcome

Recurring workflows and queued prompts run on time from your instance, even with every screen closed.

  1. 1

    Save a workflow with its schedule and branches.

  2. 2

    Let each run open its own sessions to read later.

  3. 3

    Queue the next prompt for the hour you choose.

Build & verify

Follow execution, inspect evidence and connect planned work to isolated agent runs.

06

Build & verify

Run a feature plan with an agent

Break a feature into ordered tasks in a Focus Plan, start a session that carries it, and watch the agent mark each task done with a note.
Use this when

A feature takes several ordered steps and you want progress you can check at a glance.

Outcome

Every task keeps its status and a note of what changed, and the plan closes only when you decide.

  1. 1

    Write the feature as ordered tasks.

  2. 2

    Start a session that carries the plan.

  3. 3

    Read each task’s note, then complete the plan.

07

Build & verify

Follow live agent work

Watch the plan, commands and test result update, then inspect token, time and cost statistics split by question.
Use this when

You need to understand what an agent is doing without losing the task-level thread.

Outcome

Commands, results and question-level consumption stay attached to the work they explain.

  1. 1

    Follow readable task progress as it happens.

  2. 2

    Open commands and their output without leaving the session.

  3. 3

    Review runtime, tokens and cost by question.

08

Build & verify

Review every change in Git

Open the real Source control surface, inspect a side-by-side working-tree diff, then move from implementation to test.
Use this when

An agent says the work is complete and you need repository evidence before accepting it.

Outcome

Implementation, tests and exact line changes can be reviewed beside the originating session.

  1. 1

    Open the repository working-tree status.

  2. 2

    Inspect additions and deletions side by side.

  3. 3

    Move directly between implementation and test diffs.

Understand & reuse

Measure where resources go, check each project’s health and turn verified history into reusable agent context.

09

Understand & reuse

Understand token consumption

Track tokens, sessions, runtime and API-equivalent cost, then split usage by project, app, provider and task.
Use this when

You need to explain consumption, compare projects or find an expensive workflow.

Outcome

Usage is attributable from the organization view down to projects, sessions and questions.

  1. 1

    Read the total token, session, runtime and cost picture.

  2. 2

    Compare consumption across projects and clients.

  3. 3

    Drill into providers, dates, sessions and questions.

11

Understand & reuse

Check a project’s health

Ask what is missing, stale or risky before agents work on a repository. The quick scan is instant, read-only and spends no model usage.
Use this when

You are about to hand a repository to agents and want to know what will trip them up.

Outcome

Findings grouped by domain with their severity and evidence, and every fix left as a suggestion you approve.

  1. 1

    Run the quick scan on the project.

  2. 2

    Read findings grouped by domain and severity.

  3. 3

    Open the evidence before approving a fix.

Build & distribute

Turn one sentence into a running app, then publish it for others to try and install.

12

Build & distribute

Turn one sentence into a running app

Describe the app you want and name it. An agent builds it in a new project while the live preview runs beside the conversation.
Use this when

You have an idea for an app and want to see it running before you write any code yourself.

Outcome

A new project with a running app you can try in the preview, then keep improving in the same conversation.

  1. 1

    Describe the app in one sentence.

  2. 2

    Name the app and its project.

  3. 3

    Watch the agent build it, then try it live.

13

Build & distribute

Build and publish an app people can install

Create the app with a coding agent, prepare it for phones and tablets, generate its Showcase, then submit the reviewed listing to AirCode Ø Marketplace.
Use this when

You want people to try an app in their browser and add it to their Home Screen without preparing a separate Store release.

Outcome

A review-ready Marketplace listing connected to the live Application, with Try and Install available after approval while the app is online.

  1. 1

    Build and verify the app, then run it as a saved Application.

  2. 2

    Give it a Home Screen name and icon, and generate its Showcase.

  3. 3

    Review the listing and submit the exact revision to AirCode Ø Marketplace.

Try the workflow

Give every agent a persistent place to work.

Launch from the screen you have. Continue on the screen you want.