Open the tabs you already use.
Orbital starts where the work already lives: real browser tabs, real AI tools, real windows.
Orbitalcoordinated run
Tab AChatGPT
Research angle
Tab BClaude
Code-review slice
Tab CGemini
Campaign pass
Tab DBrowser
Synthesis window
Orbital Agents
Parallel agents.Real browser tabs.
Orbital lets you spin up browser-native AI agents directly inside tabs and windows. Each one gets its own task, its own model, and its own context, so heavy work can run at once.
The product story
Not one chat pretending to be a team.
The promise of agentic work is collaboration. Orbital makes that promise literal: multiple browser windows, multiple models, multiple slices of work, one coordinated run.
Place an agent in each window.
Each agent gets its own task, model, and context, so the work can move in parallel instead of queueing behind one chat.
Let the slices run at the same time.
Research passes, review passes, and generation passes can unfold across tabs while you move to the next decision.
Bring the pieces back together.
Orbital is about coordinated workflow, not replacing the models you use. It keeps the browser as the operating surface.
Agent constellation
Every tab keeps its own gravity.
A task in ChatGPT does not have to wait for a task in Claude. A research pass does not have to share context with a campaign pass. Orbital turns the browser into a coordinated work surface.
Workflow
Tab A
ChatGPT
Research angle
own task / own model / own contextTab B
Claude
Code-review slice
own task / own model / own contextTab C
Gemini
Campaign pass
own task / own model / own contextTab D
Browser
Synthesis window
own task / own model / own contextParallel work
Heavy tasks become lanes.
The factual use cases are simple and sharp: research one topic across five sources, review several files in parallel, or generate campaign variations through multiple model perspectives.
five angles
Research brief
Run the same topic through multiple source tabs and let each agent own one angle before the synthesis pass.
Source scanAngle extractionBrief synthesis
four files
Code review
Split a diff into file slices, keep each review context clean, then bring the risk notes back into one readout.
File sliceRisk passReview handoff
ten variations
Campaign pass
Send one brief through several model perspectives and compare the strongest campaign directions side by side.
BriefVariationsSelection
Multi-model coordination
Use the right model for the right slice.
Orbital does not replace the AIs. It coordinates them through the browser. The point is not another platform; it is getting real agentic workflow out of the tools already open.
ChatGPT
Fast ideation, structured passes, familiar tabs.
Claude
Longer review work, writing passes, code context.
Gemini
Alternate model perspective for research and creative variation.
Browser-native by design
No new platform. No prompt vault. No API-key chore.
Orbital lives in the browser and follows the same local-first posture as the rest of the studio: use the tabs you have, keep the workflow close, and avoid putting a new cloud surface between you and your work.
Browser boundary
01Chrome beta
02Free at launch
03Uses the tabs you have open
04No new platform lock-in
05No API-key babysitting
06Local-first workflow posture
Core features
The page stays grounded in the product.
One window, one mission.
Each browser window can carry its own agent, task, model, and context, so agents do not collide or wait on each other.
Pick the model for the step.
Use ChatGPT, Claude, and Gemini as the work demands. Orbital coordinates the workflow around the tabs you choose.
No new platform to learn.
The agents live inside the browser surface, with the Chrome beta built around existing accounts and existing tabs.
No prompt server in the middle.
The product posture is clear: no API keys to babysit, no cloud workflow vault to trust, and no prompt collection.
Chrome beta
Let the tabs work.
Start with a research pass, a review pass, or a campaign pass. Give each window a slice and let Orbital coordinate the run.
Next coordinates