hAIve Mix
More AIs compared
Free · In‑App Purchases · Designed for iPad. Not verified for macOS.
Ask once and several AIs answer together: hAIve Mix blends them into one answer, shows you where they disagree, and checks the key claims on the web.
A single AI can be wrong — and still sound convincing. hAIve Mix queries multiple AIs in parallel, compares their answers and blends them into one synthesis, showing you where they agree and where they don't.
HOW IT WORKS
• Ask a question: up to 5 AIs answer in parallel, including ChatGPT, Gemini, Claude, DeepSeek and Mistral.
• The mixer blends the answers into one clear synthesis and highlights real divergences: if one model disagrees, you see it.
• Key claims can be verified on the web, with linked sources.
• Critical check: the AIs' consensus is stress-tested with an adversarial fact-check on important claims.
DOCUMENTS
• Attach PDFs (scans included), Word documents and spreadsheets: all the AIs analyse them together.
• Export results to Word, PDF or Excel on higher plans.
IMAGES
• Generate images with several models in parallel and pick the best one.
• Edit images with natural-language instructions.
ON YOUR COMPUTER TOO
• Same account, same history: use hAIve Mix from your browser at ask.haivemix.app.
PLANS
• Start free, with a daily question limit.
• Subscriptions increase how many AIs answer, which models are used, and the monthly fair-use limits for documents and images. Prices and details are shown in the app.
TRANSPARENCY
AI-generated answers can contain mistakes: hAIve Mix reduces that risk by cross-checking multiple models, but does not eliminate it. Always verify important information. Terms of service and privacy: https://haivemix.app/terms
more - Free plan: when a question goes over the limit, instead of "Limit reached" a card explains what happened and when you can resume.
- The Free pause is kept even if you close and reopen the app, and at the end of the month it shows the date you resume.
- During the pause "Retry" waits instead of failing.
- Small fixes and improvements.
0.1.96 6d ago
- Free plan: when your questions run out, a card in the chat shows when you can resume and what the Minimal plan offers, with no interrupting pop-ups.
- During the pause you can already type your next question: the field tells you from what time you can send it.
- Light theme easier to read on some buttons and in the Add-ons section of Settings.
- Examples on the home screen: "Try it" shows which plan you need to redo the example.
- Small fixes and improvements.
0.1.95 Oct 24
- More reliable dictation: when you send with the microphone on, the app waits long enough not to lose the last words.
- Marks in the synthesis cover the whole sentence instead of stopping halfway.
- If the synthesis fails, a clear note with an icon explains what happened.
- Intro: plan prices are those of your country, in your currency.
- Small fixes and improvements.
0.1.94 Oct 23
- New Guide: open it from "How hAIve Mix works" on the home screen to see synthesis, checks, documents and images step by step, with the AIs in use.
- Examples on the home screen: sample chats you open with a tap to see what each plan does, and "Try it" to redo them with your own questions.
- Reorganised Settings: your account with plan and usage at the top, a clearer "Extend your limits", notifications and privacy in their own sections.
- Refined plans page, with what each plan includes made clearer.
- Usage limit shown on the send button; when Free plan questions run out you see when you can resume and get a reminder.
- Redesigned disagreements sheet: the AIs' values and ranges compared on a scale.
- Faster intro, with the new "5 AIs, one app" screen.
- Images: fixed some cases where a finished image didn't appear.
- Small fixes and improvements.
0.1.93 Oct 22
- Redesigned check sheet: the verdict stands out in a coloured band, with the logos of the AIs involved and the statement checked.
- Check sources now also come in the language of the answer, with a language tag when it differs from the app's.
- AI logos also appear in the panels under the synthesis and no longer break across lines.
- Smoother chat while scrolling through long answers.
- Small fixes and improvements.
0.1.92 Oct 16
- New Minimal and Minimal+ plans for people who use hAIve Mix every day, with a switch to add DeepSeek.
- Redesigned plans page: the same items on every card, showing what's included and what the next plan unlocks.
- Choose your look: Classic plus six colour schemes, each with its own light and dark version, with a preview.
- Cleaner synthesis: marks are light underlines, with the AI logos before each marked statement and a legend at the bottom.
- Tables in the synthesis are now real tables, easier to read.
- Images: if a generation fails, the reason is shown under the slot.
- Dictation: when you send with the microphone on, the last words are no longer lost.
- Small fixes and improvements.
0.1.91 Oct 14
- Sign-in: Google is available on iPhone too, next to Apple, with the two buttons side by side.
- On launch your plan is already the right one: no more "Free" flash or trial banner for returning users.
- Signing out and deleting the account always take you to the sign-in screen.
- Redesigned intro: four screens with less text; theme and language are chosen right away.
- Images: the image you're working on stays active throughout the conversation.
- Purchases: if the store doesn't respond, the app retries and lets you know.
- Small fixes and improvements.
0.1.90 Oct 12
- New glass-style writing bar, in both dark and light themes.
- Home cards easier to read in the light theme.
- Settings: "Tips and offers" is now in the Notifications section.
- Small fixes and improvements.
0.1.88 Oct 2
- While the synthesis is being prepared you can watch the answers coming in: the AI logos show who is still writing and who has finished.
- Web check: clearer progress while it runs, and a "Try again with a new check" button under a verdict already given.
- New "Report an inaccuracy" to tell us about a wrong answer.
- Reminders during the Max trial, so you don't forget when it ends.
- Small fixes and improvements.
0.1.87 Oct 1
- At the bottom of the synthesis you'll find "Go to the individual model answers"; each answer has "Close answer" and the "Read" buttons open the model's answer directly.
- When the synthesis is low-confidence, quick buttons take you to the model answers; on creative requests, cards show the start of each proposal.
- Interface now available in Korean.
- Read aloud: the voice follows the language of the text and tells you when that voice isn't available.
- The count of answering AIs no longer includes those temporarily unavailable.
- Small fixes and improvements.
0.1.86 Sep 30
- Refined home cards: AI logos visible on iPhone too, the active card highlights the name in the AI's colour, wider question field.
- Documents: tapping a document lets you choose between Open and Share or save.
- Plans: the fair-use terms are at the bottom of the page and the plan card no longer scrolls on its own.
- Small fixes and improvements.
0.1.85 Sep 29
• New home: the AIs are cards above the question field. Tap to switch one on or off, long-press to get an answer from a single AI; below you see how many AIs your question goes to.
• Refreshed swarm animation, now with the AI logos.
• Annotated copy of documents: in Word the summary of comments is at the end of the file; OpenDocument (.odt) files are supported too.
• Small fixes and improvements.
0.1.84 Sep 28
- Refreshed home screen: new animation of the AIs in orbit and a redesigned Max trial banner, noting that when it ends you go back to the Free plan with no card and no charges.
- The Max trial is offered at precise moments: on the home screen, under the first answer and under the seventh.
- In the side menu, "Send a suggestion / problem" to write to us directly; every report is read.
- With many conversations, "New chat" becomes a "+" button next to the search.
0.1.83 Sep 28
- When the free plan, the trial or your plan period runs out, a clear card under the last answer shows the time left and what you can do.
- Fixed a purchase reported as failed when the store flagged a double tap: only one purchase flow at a time across the app now.
- Verification highlights no longer overlap each other or touch links in the answer.
0.1.81 Sep 25
- Attached photos are all kept: a message with several photos shows them all when reopened, and they are no longer shrunk.
- Updated image limits: 4 trial generations on the free plan, 6 image turns in the Max trial; Go and Pro monthly limits go down by 5.
0.1.80 Sep 24
- Fixed dictation repeating the words you had just said, with the text growing in duplicated pieces.
- The reliability estimate no longer says "high" when the AIs actually disagree.
- The "How hAIve Mix works" card shows its steps even without scrolling.
0.1.79 Sep 19
- When every AI refuses an image, the answer says which ones said no and suggests the more permissive models.
- Minority-position pills are easier to read: bold text and map links no longer break the text.
- When you verify a pill, the address inside stays tappable.
0.1.78 Sep 15
- Fixed purchases failing right after opening the app: the app now waits for the store to be ready, and says precisely what went wrong if something does.
- Conversation sync no longer gets stuck in some cases after switching account.
0.1.76 Sep 15
- A deleted conversation no longer comes back when you open the app on another device.
- Documents are read more faithfully, tables included.
0.1.75 Sep 13
- Answers you deleted or replaced (edit and resend, retry) no longer show up twice when you reopen the conversation from another device.
- Attached documents are read more faithfully, tables included.
- On the plans screen, the Pro card lists the AIs for images and what document and photo analysis includes.
- Tapping a locked model badge on the home screen opens the right plan.
- The attached-photo notice no longer covers the text field while the keyboard is open.
0.1.74 Sep 11
Minor improvements and fixes.
0.1.73 Sep 9
- More predictable image editing: the base is the image you pick, with a "Used as base" badge on earlier turns' tiles.
- The Max trial states the image turns it includes, with a counter in Settings and the asterisk note.
- Pasted text gets cleaned up, and very long texts get a warning that states the limit clearly.
0.1.72 Sep 8
- When an image is not generated, the card states the real reason: "Subject refused by the model" or "Timed out", no more generic error.
- With a single AI active a "single answer - no comparison" pill appears, and documents and web verification work there too.
- Verification highlights anchor to the text more reliably, even when the quote wraps across lines.
- Polished image loading visuals.
0.1.71 Sep 6
- The attachment button now opens on every plan: you can pick photos and camera, and only documents stay reserved for higher plans.
- An attached photo can be used as the base for editing an image on any plan.
- Generated images stay available when you reopen a conversation, on the free plan too.
- Tapping Retry after an error no longer loses the attached photo.
- When a short wait is needed the app calls it a pause, without using up your trial.
0.1.69 Sep 3
- Answers you paid for but never received are recovered when you reopen the app: no more conversations that vanish halfway.
- Large photos are resized instead of rejected, and the size limit message now states the real limit.
- When dictation cannot start because the microphone permission is missing, the app says so instead of calling it unavailable.
- Fixes to saving exported documents with unusual titles.
0.1.67 Sep 2
- Free plan: when a question goes over the limit, instead of "Limit reached" a card explains what happened and when you can resume.
- The Free pause is kept even if you close and reopen the app, and at the end of the month it shows the date you resume.
- During the pause "Retry" waits instead of failing.
- Small fixes and improvements.
more Version 0.1.96 6d ago
Data Linked to You The following data may be collected and linked to your identity:
Purchases Contact Info User Content Search History Identifiers Usage Data Diagnostics 
