Voice · Vision · Code · Companion agent

Build, run, and iterate alongside an agent that sees your screen and hears you think.

Share your screen and camera, talk through the problem, and ship working code in the same pane — with a Companion agent that holds context across sessions.

Start building View privacy policy

Multimodal context, every turn

Live screen-share, camera, voice, and attachments fold into each prompt. Server-side speech-to-text and OCR-grade screen reading mean the model actually sees what you see.

Code, run, and iterate in one pane

A Monaco editor and sandboxed preview sit next to chat, with an agent harness that drives real runs and saves the artifacts they produce.

A Companion that sticks around

The Companion agent rides along in a side panel, remembers your project across sessions, and keeps shared threads and reusable artifacts within reach.

Simple pricing

Start free, then scale up as voice turns, vision frames, and agent runs add up.

Public500 credits / month

$0/mo

A lightweight way to try voice, screen-share, and the Companion agent on a real project.

Get started

Starter5,000 credits / month

$9/mo

For independent builders shipping side projects with regular voice + agent sessions.

Choose Starter

Pro15,000 credits / month

$19/mo

For heavier daily use — longer agent runs, more vision frames, more sandbox iteration.

Choose Pro

Premium40,000 credits / month

$39/mo

For teams and full-time usage that need the most headroom for voice, vision, and agent work.

Choose Premium