ZIMA

FEATURES

A lot more than chat.

Thirteen things ZIMA does that no single-vendor app can.

Every model, one chat

ZIMA brings 25 models from OpenAI, Anthropic, and Google together in one app. Open the model selector at the top of any chat to switch — you keep the full conversation when you do, so you can start with a fast free model and escalate to a frontier model for the hard part.

Generate images

Just describe what you want ("draw a…", "make an image of…") and the model calls the image tool automatically. Tune size, quality, and number of images from the conversation's image options.

Generate video

Ask a Gemini chat for a short video and the model calls the Veo tool. Attach an image to use it as the first frame, or attach a video to edit it.

Voice chat

Tap the voice button to enter a minimal voice mode. Every voice model can search the web mid-call, you can attach a document or image while talking, and on mobile the conversation keeps going if you lock the screen. With Gemini you can even share your live camera feed. Everything is transcribed into your chat history.

Language translation

Ask how to say something and you get more than a one-line translation — the phrase itself, how to pronounce it, and a piece-by-piece breakdown of what each word does, so you can actually use it with confidence.

Live web search

There's no toggle to flip — just ask. When a question needs live or current information, the model searches automatically and lists the sources it used beneath the reply. Tap any source to open it.

Find real photos on the web

Say "find me photos of itasha cars" and ZIMA opens an image site, downloads the real photos, and shows them in a card with thumbnails and the source. Works with every model. The photos stay in the conversation — ask about them, or drop one into a document ZIMA is designing for you.

Extended thinking

For models that support it (the GPT-5.6 and GPT-5 families, Claude Fable/Opus/Sonnet, Gemini 2.5/3.x), ZIMA can let the model reason longer before answering — better results on complex, multi-step problems.

Image & document uploads

Attach up to 4 items per message (images plus documents, 50 MB each). Claude and Gemini read PDFs natively, preserving layout and images; other file types are extracted to text. Supported docs: PDF, .txt, .md, .csv, .docx, .xlsx.

Google Drive

Ask "find my Q3 budget sheet and summarize it" and ZIMA searches your Drive, reads the file, and shows a card with exactly what it read. It creates Docs, Sheets, Slides, and Forms in your Drive, and you can attach up to 3 Drive files per message from the + menu — linked, not copied. The only files ZIMA modifies are ones it created or copied itself. Rolling out now.

Gmail

Ask about an email and ZIMA finds the thread, reads it, and shows a card with exactly what it read. Ask for a reply and it writes a Gmail draft — drafts are never sent; you review and send them yourself. ZIMA can never send email or delete anything. Rolling out now.

Create real files

ZIMA creates Word docs, PDFs, Excel workbooks (formulas included), PowerPoint decks, CSV, Markdown, JSON, HTML — even ZIP bundles — up to 5 files per message, with any chat model. A progress card settles into a file card with Preview and Download. To change a file, just ask ("make the intro shorter") and ZIMA updates the same file. Files are kept for 30 days — download them or save them to Google Drive to keep them permanently.

Synced history

Conversations sync across your signed-in devices, so you can start on your phone and pick up on the web. Rename or delete chats from the sidebar. (An Android app is in development.)

25+
models in one app
3
frontier labs
$0
to start

Try all of it. Free to start.

No sign-up to start · iOS & web

Start chatting