Apps from a conversation
Ask for a tracker, a tool, a little game — your assistant builds it and shows you the running result.
Two modes, one product. Basic is the everyday assistant — chat, your apps, and the Store. Same app, same account.
Two modes, one product. Advanced opens the full deck — instances, workspaces, and orchestrators. Same app, same account.
Powered by WormholeWarpdrive gives you a personal assistant that builds real apps from a plain-language conversation. Install more from the Store with one tap, and run source-based Apps in a local Instance on your own computer.
Warpdrive runs your agentic coding tools — Claude Code, Codex, Copilot, or agents you build — in separate Instances on hardware you choose. Launch from your desktop or your phone.
Tell your Warpdrive assistant what you need — a tracker, a planner, a tool nobody makes. It builds it in front of you, and it’s yours, on your own computer.
A chore chart, a garden log, that one tracker no app store has. Ask your assistant and it makes it in a local Instance on your own computer — no coding involved.
Choose where work runs and which AI it uses. Keep source-based Apps and Local AI on hardware you control, with a clear account of what every connected service receives.
Scan a QR code and your phone joins your Warpdrive. Chat with your assistant, open your apps, and check progress from the couch — anywhere on your network.
The Warpdrive Store is opening to founding creators — verified builds, GitHub import, and a payouts pipeline ready from day one. Get in before the doors open.
Claude Code, Codex, Copilot, OpenCode, or your own — one isolated sandbox per agent, with chat, terminal, files, and live previews in a single workbench you can open from any device.
Queue tasks and let Conductors dispatch, run, and retry work across your agents. Auto-Tune benchmarks every change and keeps only measurable wins.
Put Warpdrive on hardware you control, keep your projects close, and use the same command center from your desk or your phone.
Everything you need to make apps of your own — no coding, no setup maze.
Ask for a tracker, a tool, a little game — your assistant builds it and shows you the running result.
Everything you build or install lives in Apps — tap one and it opens full-screen, running live.
Open an app and tell your assistant what you want different — it updates right where you are.
Every project is a thread you can come back to — star the important ones, or just talk instead of typing.
Browse ready-made apps and tap Get — your assistant sets everything up for you.
Scan a QR code and your phone joins in — assistant, apps, and conversations, anywhere on your network.
One command deck for your agents — from first launch to landed mission, on machines you own.
Claude Code, Codex, Copilot, and OpenCode ship in the box — sign in to each one right inside its sandbox.
Any CLI agent, script, or MCP server you install runs the same way — your own tooling, same isolation.
Give each project a clean place to work, with its own tools and access, so one mission does not spill into another.
Conductors queue tasks, dispatch agents, and retry failures — Auto-Tune keeps only measurable wins.
Chat, workspace files, a live terminal, and app previews — side by side in the same view.
The same app on desktop, web, tablet, and phone. Scan a QR code and your phone joins your Warpdrive.
Install verified, exact-revision builds — with creator payouts and GitHub import opening soon.
Take a snapshot before a risky run and restore the workspace if the agent goes sideways.
Sign in to the CLIs you already pay for, or drop in an OpenAI, Anthropic, or Venice key — even a local model.
Use the laptops, desktops, and home servers you already own as capacity for more work.
One request becomes a working app you can open on desktop or phone.
Your assistant starts with the outcome, not a setup form.
The assistant plans the app, creates its files, and checks the result.
The finished Chore Tracker appears beside the apps you already use.
Assign chores, mark work done, and see the household move forward.
Open Chore Tracker away from your desk. The app and its progress stay with you.
Instances do the focused work. A Conductor keeps the mission in step.
Each Instance has a clear job, machine, and live state.
The task arrives with its goal and project context attached.
Chat, files, terminal, and preview stay together while the agent builds.
The Conductor routes tasks, waits for checks, and brings the agents’ results together.
Ready to configure a machine? Follow Getting Started in the documentation.
Short answers about AI, privacy, pricing, Apps, and where your work runs. Technical details link to the full guides.
Warpdrive is an AI assistant and app builder. Basic keeps it simple; Advanced gives you full control of agents, Apps, and where they run.
No. In Basic, describe what you want in plain language. Advanced tools are there when you want to inspect or control the technical details.
On an Instance you choose — usually your own computer, another paired machine, or Warpdrive Cloud. Learn how Instances work.
Your Apps, Conversations, and App data can stay on your hardware. A cloud AI provider receives the prompts you send to it; a local model keeps that work local. Read the privacy details.
Yes. Warpdrive can install a supported model on your computer, or connect to an existing local endpoint. See the Local AI guide.
Yes. Use supported providers such as Codex, Claude Code, Copilot, and OpenCode, or connect a compatible local model. You sign in when its Instance starts.
Basic centers on your Assistant, Store, Apps, and Conversations. Advanced adds Instances, Workspaces, Orchestrators, Conductors, terminals, and diffs.
Yes. Pair your phone with a QR code to chat, open Apps, and check progress on your local network or through a hosted relay.
Apps and AI running on that computer pause until it is back online. Work on Warpdrive Cloud or another running Instance stays available.
Instances are isolated work areas where Apps and AI tasks run. Each Instance has its own location, Workspace, tools, and status. Read the Instances guide.
Conductors coordinate the agents in an Orchestrator. They manage the run and give you one Conversation for progress, questions, and results.
Warpdrive is free to self-host. Paid AI providers may charge for model use, and Warpdrive Cloud uses a free trial followed by a subscription.
Yes. Share an App through your reachable Gateway. Store publishing is rolling out to founding creators and appears when it is enabled for your deployment.
The Store is the catalog where you discover and get Apps. Use one as-is or make your own version with your Assistant.
Choose the visual app when context matters. Choose the terminal-first CLI when speed and scripts matter. Use both when the job crosses machines.
Sign in and you get the whole deck: instances, workspaces, orchestrators, snapshots, diffs, and the Store, with the same work close at hand on every screen.
A terminal-first way to create and operate the same work. It fits into the shell habits and automation you already use.
Most people use both. Keep the visual app where you plan and review. Put the CLI on machines where work runs. The documentation explains the setup, requirements, and remote connection flow. Read the CLI documentation.
Install the app on your computer and sign in. That is the whole setup — the rest is optional.
Start with the experience you want. Add the CLI or mobile companion when your work grows across machines and screens.
The full desktop experience
Run your own Warpdrive on your computer and save on fees.
Want it free? Open Settings, pick a Local AI model matched to your machine, and it installs itself — no API key, nothing per question.
The terminal-first experience
Run focused AI work from the shell, automate repeatable jobs, and add capacity on the machines you already own.
Everything, in your pocket
Scan a QR code and watch your apps come to life from anywhere.
Your computer keeps doing the work back home.
Need another platform or installation path? See every build in the documentation.