whipcode
Use cases

What you can do with Whipcode.

You already run the agents. These are the everyday jobs Whipcode is built for, so you spend your attention on the work that actually needs you.

Supervise

Run several agents without living in tabs.

One agent is editing. One is stuck. Another finished ten minutes ago. Finding out shouldn't mean opening every terminal.

Whipcode brings those conversations into one workspace. You see who is working, who is waiting and who is done at a glance, then open the conversation that needs you. Each conversation streams live: text, tool calls, subagents and diffs.

  • One sidebar for every conversation
  • Live status: working, needs you, finished
  • Switch the model per conversation
Remote control

Manage Codex and Claude Code across machines.

Your repos and CLI logins live on specific machines: the Mac at your desk, a Linux box, a server. Whipcode runs the agents where they already belong and brings the control to one browser.

A client on each machine connects out to your hub, so nothing needs a public IP or an open port. You steer Codex on the server and Claude Code on the laptop from the same place, and send instructions or answers to either without switching machines.

  • Claude Code, Codex, Grok, Gemini, Kimi, Muse
  • Agents run on the machine with your repo
  • No inbound port on your machines
From your phone

Answer a blocking question away from your desk.

An agent waiting for an answer is an agent doing nothing. When one needs a decision, the question shows up as a card you can answer, even from a mobile browser built for a smaller screen.

Reply to the Codex question that's holding up a fix, or tell Claude Code to keep guest checkout, without reopening your laptop. The agent keeps running on the connected machine, which has to stay online and awake; the phone is just how you reach it.

  • Questions as answerable cards
  • Mobile workspace, add to home screen
  • Same conversations, a smaller screen
Delegate

Hand a multi-step job to a mission.

Some work takes more than one turn. A mission lets an agent keep going toward a goal without waiting for your next prompt, while a separate AI judge checks the result and sends it back for another pass when something is missing.

Missions run within configured limits and can hit a limit, get blocked or reach a decision that needs you. You inspect progress, pause or stop. The judge is an extra check, not a guarantee: it doesn't replace testing and your review.

  • Goal-oriented across turns
  • AI-judged, watched for stalls
  • Pause or stop anytime
Keep your terminal

Drop back to the CLI whenever you want.

Whipcode sits beside your terminal, not on top of it. Claude conversations use the CLI's own session files, so you can pick one up on the machine with claude --resume <session-id>.

Classic tmux terminals, one isolated socket per project, are reachable from the hub when a job calls for a plain shell. Session continuity depends on the engine, so resume isn't a promise for every provider.

  • Resume Claude sessions in the CLI
  • tmux terminals from the hub
  • Attach a screenshot or file for context
Orchestrate

Direct more than one conversation at once.

SuperWhip turns one plain-language instruction into a reviewed plan of actions across several conversations. You read the plan before it runs, so the fan-out stays under your control.

Prefer the machine to pick? Auto routes a prompt using the available models and your engine preferences. An optional voice mode and the Whipcode MCP server give connected AI clients another way to work with your conversations and missions.

  • SuperWhip: one instruction, many conversations
  • Auto engine routing
  • MCP server and opt-in voice

One interface for all your engines.

  • Claude Code
  • Codex
  • Grok
  • Gemini
  • Kimi
  • Muse
  • Auto

More room to build. Less to juggle.

See how the pieces fit together, or request access to the private beta.

Get early access How it works → Private beta · Engine usage billed by your providers.