PRODUCT TOUR

Every part of the window,
explained.

An interactive tour of the real app with example projects. Pick a pin to see how that part works.

Download free trial See pricing

Works with

Claude CodeCodexOpenCodeAntigravity

Codex drafts and Claude reviews together in Office mode.

Review branches, commit history, changed files and a selected diff in the Git window.

Three mobile views: task control, the shared Office and Controller mode with project and task selection and voice-responsive push-to-talk.

The Office zoomed onto one agent: its conversation opens beside the desk, with a switch to its terminal. Project tabs, the queue, the new task composer and all tasks across projects stay around it.

The VibeiDE Office across the whole window, with the queue hidden and the new task composer floating over the room.

Six agents in the Grid across the whole window, with the queue, composer and bottom task bar hidden.

Plan council

One agent drafts, the other challenges, the author revises. Any two providers work, including OpenCode and Antigravity.

Office view

Every running task is an agent at its own desk, labeled with its task, model, effort and time. Click a desk to zoom in and open that agent as a chat or as its live terminal.

Brain

Turn on the Brain and your agents keep your corrections and what they learn, and use it in later tasks. It is off by default: one switch, nothing to set up.

Scripts

VibeiDE finds every npm script in the project, nested packages included. Each run opens in a Terminal agent you can watch, stop and run again.

Usage

Each limit is measured against the time already gone in its window, and VibeiDE says when it runs out at your current tempo.

Review

Finished tasks wait in their own column with the full terminal and chat history. Open one, read the diff, reply or mark it complete.

Terminal grid

Put up to 20 task terminals side by side. Switch any pane between the raw terminal and the conversation, or hide the queue to give the panes the room.

Local voice input

Hold your shortcut or the microphone in any composer to dictate a prompt or follow-up. A speech model on your own computer turns what you say into text, so no cloud service hears it.

All tasks

The bar along the bottom shows work from every project: running, waiting for you, ready for review or queued. Click a card to jump to that task.

Branches

Browse branches, commit history and changed files in the Git window. Two Git modes, chosen per project, keep parallel edits apart.

Review

Finished tasks wait in their own column with the full terminal and chat history. Open one, read the diff, reply or mark it complete.

Mobile

Pair with a QR code, then queue, watch and review from anywhere while your computer does the work.

Push to talk

Hold the microphone on your paired phone, speak and release. Your computer transcribes it locally and you check the text before sending.

Projects

Keep several repositories open as project tabs along the top. Each one has its own queue and its own limit of parallel tasks for every provider.

Review

Finished tasks wait in their own column with the full terminal and chat history. Open one, read the diff, reply or mark it complete.

Project queues

Free slots pick up the next request, and a waiting task never blocks the ones behind it.

Antislop design

Turn on Antislop design in Options for interface work. The agent starts from your design library, reuses what is there and adds the components it builds.

Office view

Every running task is an agent at its own desk, labeled with its task, model, effort and time. Click a desk to zoom in and open that agent as a chat or as its live terminal.

Brain

Turn on the Brain and your agents keep your corrections and what they learn, and use it in later tasks. It is off by default: one switch, nothing to set up.

Chat or terminal

Open the zoomed agent as a chat or as its live terminal, then type to it.

Local voice input

Hold your shortcut or the microphone in any composer to dictate a prompt or follow-up. A speech model on your own computer turns what you say into text, so no cloud service hears it.

All tasks

The bar along the bottom shows work from every project: running, waiting for you, ready for review or queued. Click a card to jump to that task.

Turbo mode

A planner splits your task into smaller parts. Then every part runs at the same time on its own agent, as many agents as you allow in the project.

Levels

Finish tasks, earn XP and build your dream office. Every level brings an upgrade.

Office view

Every running task is an agent at its own desk, labeled with its task, model, effort and time. Click a desk to zoom in and open that agent as a chat or as its live terminal.

Brain

Turn on the Brain and your agents keep your corrections and what they learn, and use it in later tasks. It is off by default: one switch, nothing to set up.

Scripts

VibeiDE finds every npm script in the project, nested packages included. Each run opens in a Terminal agent you can watch, stop and run again.

Terminal grid

Put up to 20 task terminals side by side. Switch any pane between the raw terminal and the conversation, or hide the queue to give the panes the room.

Review

Finished tasks wait in their own column with the full terminal and chat history. Open one, read the diff, reply or mark it complete.

20task terminals side by side
4agents: Claude Code, Codex, OpenCode, Antigravity
1queue and limit per project
0open ports for mobile control

QUEUES, OFFICE AND GRID

One queue per project. Every agent in view.

Each repository gets its own tab, queue and parallel-task limit per provider. Watch them as desks, as a grid, or all at once.

Project queues

Every repository gets its own tab, queue and parallel limit per provider. Tasks wait their turn instead of fighting for the same terminal.

How task queues work

Office view

Running tasks show up as agents at desks. Click a desk to open its chat or live terminal.

❯ task #4460
… running
edit src/api.ts

❯ task #4461
✓ tests pass

❯ task #4462
✓ tests pass
edit src/api.ts

❯ task #4463
… running

❯ task #4464
✓ tests pass
edit src/api.ts

❯ task #4465
✓ tests pass

❯ task #4466
… running
edit src/api.ts

❯ task #4467
✓ tests pass

❯ task #4468
✓ tests pass
edit src/api.ts

❯ task #4469
… running

❯ task #4470
✓ tests pass
edit src/api.ts

❯ task #4471
✓ tests pass

Terminal grid

Up to 20 task terminals side by side when you want to watch everything at once.

All tasks, every project

The bottom bar shows running, ready and queued work across every project, grouped the way you like.

SESSIONS

Every task starts clean and can be resumed.

A fresh session per task keeps context small, cuts hallucinations and saves tokens. Pick any task up again later.

Fresh context for every task
Resume any finished task later
Smaller prompts, lower token cost

Context size per turn

One long session

One per task

In one long session the context size grows with every turn. With one session per task it stays small and repeats.

Illustration, not measured data.

REVIEW

Finished work waits for your review.

Every task keeps its own terminal and history, so you review results one by one, not in a wall of scrollback.

Its own terminal and chat per task
Diff and history in one place
Reply, retry or mark complete

PLAN COUNCIL

Two agents check the plan before code exists.

One agent drafts, another challenges, the author revises. Any two providers work, including OpenCode and Antigravity.

Step 1
Claude writes Drafts the plan from your request.
Step 2
Codex checks Challenges gaps and risky steps.
Step 3
Claude revises Answers each point and tightens it.
Step 4
You hand off Approve and send it to the queue.

BRAIN Off by default

Agents that remember your project.

01
Reads first

Before a task starts, the agent reads the lessons that apply.
02
Saves last

When it finishes, it saves what it learned, briefly.
03
Stays true

Lessons cite code and pause when that code changes.

Plug and play
No chores
Small
Local
Light

Built after studying Mem0, Graphiti, Hindsight, Basic Memory, Claude-Mem and Agent Beacon, plus more than 1,600 of our own sessions.

DAILY TOOLS

The tools around the queue.

Mobile

Your queue, in your pocket. Pair by QR code; outbound HTTPS only, no open port.

Push to talk

Hold. Speak. Release. Local speech model, 670 MB, no GPU.

Turbo mode

One plan, many agents. A planner splits work into parallel sessions.

Changelog

Turn finished work into a changelog, grouped by category.

Design gallery

One reusable design system in one folder, with a 0–100 audit.

Antislop design

Builds your design library and skips 68 common AI-look patterns.

Scripts

Every npm script one click away, nested packages too.

Usage

See a Claude or Codex limit run out before it does.

Worktrees

Keep parallel edits apart with a worktree per task.

Levels

A gamified office with rewards and achievements.

COMPANY MODE · EARLY ACCESS

Your team. One company view.

Shared task names, activity and review work. Prompts, code and output stay private unless members share them.

Explore Company mode

DOWNLOAD

Download VibeiDE

14-day trial, no card. Works with the Claude Code, Codex and OpenCode you already have.

Stable

macOS

DMG for Apple silicon and Intel

Download for macOS

Stable

Windows

Setup installer. SmartScreen may ask once.

Download for Windows

Stable

Linux

AppImage, runs on most distros

Download for Linux

FAQ

Questions, answered.

Why not one Claude Code session?

One long session keeps growing until it slows down and forgets. VibeiDE gives every task its own clean session and a queue, so many agents work in parallel without stepping on each other.

Why not tmux or separate terminals?

tmux and separate terminals work if you track everything yourself. VibeiDE adds project queues, provider limits, saved history and review, so no result gets lost in a tab.

How many tasks can run at once?

You choose. Each project has its own limit of parallel tasks for every provider, such as Codex, Claude Code and OpenCode. When a limit is reached, new tasks wait in the queue and start as soon as a slot frees up.

Do I need new API keys?

No. VibeiDE runs the Codex and Claude Code CLIs already on your computer. Your logins, keys and credits stay with them.

Does VibeiDE include Claude Code or Codex usage?

No. VibeiDE runs the Codex, Claude Code or OpenCode you already use, with your own subscription or API key. Your provider bills its own usage and applies its own limits. One provider is enough to start.

Which computers does it run on?

VibeiDE runs on macOS, Windows and Linux. Get the installer for your system from the download page.

Can I cancel?

Yes. Cancel any time in Billing. Your plan stops renewing and access continues until the end of the paid period. You add no payment card for the 14-day trial, so nothing renews after it.

Can I keep a terminal open between turns?

Yes. Turn on Keep open when you queue a task, then finish it when you are done.

What happens to lifetime licences if VibeiDE is discontinued?

Your licence to the last supplied version remains. Updates and hosted services may stop. The app currently requires periodic online licence validation. If development or hosted licensing permanently ends, we will notify affected customers and explain the available continuation, migration and statutory remedies. See the licence and sale terms.

Where is my data stored?

VibeiDE stores your task records on your computers. Your AI provider receives prompts and context under its own terms. With current versions, Company mode, switching between and downloading tasks from your own computers, and optional Mobile control are end-to-end encrypted: the data goes between your devices directly or through VibeiDE's relay, with keys that only your devices hold, so VibeiDE's servers can pass it along but cannot read or store it. Bug reports you choose to send, with any screenshots, are stored with your account until you delete them. See Privacy.

Every part of the window,
explained.

One queue per project. Every agent in view.

Project queues

Office view

Terminal grid

All tasks, every project

Every task starts clean and can be resumed.

Finished work waits for your review.

Two agents check the plan before code exists.

Agents that remember your project.

Reads first

Saves last

Stays true

The tools around the queue.

Mobile

Push to talk

Turbo mode

Changelog

Design gallery

Antislop design

Scripts

Usage

Worktrees

Levels

Your team. One company view.

Download VibeiDE

macOS

Windows

Linux

Questions, answered.

Find your workflow

Tools and teams

Use cases

Comparisons

Every part of the window,explained.

One queue per project. Every agent in view.

Project queues

Office view

Terminal grid

All tasks, every project

Every task starts clean and can be resumed.

Finished work waits for your review.

Two agents check the plan before code exists.

Agents that remember your project.

Reads first

Saves last

Stays true

The tools around the queue.

Your team. One company view.

Download VibeiDE

macOS

Windows

Linux

Questions, answered.

Find your workflow

Tools and teams

Use cases

Comparisons

Every part of the window,
explained.