Skip to content

Catalyst on iPhone

Catalyst has a native iPhone app. It signs into the same account and shows the same conversations as catalyst.voov.ai — start something on your phone, finish it at your desk.

It is deliberately not the whole web app squeezed onto a small screen. The phone is built for the things you actually do on a phone: chatting, talking, and reading answers. Building things — workflows, mini apps, data sources — stays on the desktop, where there’s room for a canvas and a code editor.

A Catalyst chat on iPhone showing a generated image of a golden eagle inside the conversation.

Full chat

Streaming replies, reasoning blocks, tool runs, code output, generated images and inline video playback — the same conversation you’d see on the web.

Several models at once

Ask the same question of multiple models and flip between the answers with the side rail.

Voice

Dictate instead of typing, or open hands-free voice mode and just talk.

Attachments

Add photos from your library, take a picture, or attach files from Files.

Providers & models

Add a service, paste an API key, browse its models and pick your default — all from the phone.

Tools

Turn built-in tools and your MCP servers’ tools on or off per chat from Add to Chat.

Pick more than one model and the reply arrives as a set. The rail down the right edge switches between them, so you can judge answers against each other on the phone the same way you would on the web.

An iPhone chat with three models answering the same question, a vertical rail on the right listing claude-fable-5, gpt-5.5 and gemini-3.1-pro.

Tap the mic to dictate into the composer, or tap the waveform for voice mode — a full-screen conversation where Catalyst listens, sends when you pause, reads the reply aloud, and reopens the mic for your next turn.

Catalyst's full-screen voice mode on iPhone: a large blue orb reacting to the voice, captioned LISTENING · APPLE, with the transcribed question beneath.

Speech recognition uses Apple’s on-device recognizer, and replies can be read by the system voice or by an optional neural voice you download once (about 350 MB) in Settings › Voice. Either way, your audio never leaves the phone.

The Voice section of iPhone Settings, offering the neural voice download, the reply voice, and a Speak replies aloud toggle.

Add providers, paste keys, and browse each service’s models without opening a laptop.

The Services screen on iPhone listing connected AI providers.

These surfaces have no iPhone equivalent yet. Everything they produce still works on the phone — a workflow you built on the desktop keeps running, a skill you wrote still shapes answers, memory the assistant saved still comes back in conversation — you just can’t build or manage them from iOS.

SurfaceOn iPhoneWhere to use it
WorkflowsBuild, run and schedule on the web
Mini appsBuild, run and publish on the web
Data sourcesConnect and query on the web
PromptsYour prompt library lives on the web
SkillsAuthor and edit on the web
MemoryUsed automaticallyReview and edit entries on the web
MCP serversTools usable per chatAdd and configure servers on the web
Sharing a chatCreate share links on the web
Chat search & historyRecent sessions onlyFull search on the web
  • iOS 18 or later, on iPhone and iPad.
  • A Catalyst account — the same one you use on the web.
  • At least one connected model provider (add it on either platform).