# Anuma - Private AI with Memory — Every AI Model. One Memory.

> Unified access to multiple AI models with persistent, private memory. (iPhone/iPad app by Anuma AI.)

- Source: https://appshunter.io/ios/app/anuma-private-ai-with-memory/id6758104125 (this page in markdown: same URL + `.md`)
- Developer: [Anuma AI](https://appshunter.io/developer/1870887678)
- Category: Productivity, Utilities
- Price: Free with in-app purchases
- Rating: 4.40/5 from 5 App Store ratings · 1 written reviews indexed
- Age rating: 4+
- Requires: iOS 17.5 · 146 MB
- Languages: American English
- Released: 2026-03-24
- Data updated: 2026-07-31
- Monetization: free with in-app purchases
- User reviews in markdown: https://appshunter.io/ios/app/anuma-private-ai-with-memory/id6758104125/reviews.md

## What is Anuma?

Anuma brings together the world’s leading frontier and open-source AI models, including Claude, ChatGPT, Gemini, Kimi and more, in a single secure interface while keeping your chat history and memory private and under your control. 

With Anuma, you can use the best AI models without managing multiple apps or losing your context. You can build your entire AI workflow in one place that you fully own.

USE THE BEST AI MODELS
Different AI models are good at different tasks. With Anuma you can access multiple leading models in one environment and switch between them without losing context. Supported models include ChatGPT, Claude, Gemini, Grok, Qwen, Ling, Kimi, MiniMax, Flux and Nano Banana.

MEMORY THAT WORKS ACROSS EVERY MODEL
Anuma Memory stores your preferences and the context behind your work. It carries across every model, every device and every session. Switch from Claude to GPT mid-conversation and nothing is lost. Start a new chat next week and your AI already knows who you are, what you're working on, and how you like your answers.  The more you use Anuma, the less you have to explain. And unlike other platforms, your memory is not locked inside one model. It belongs to you, works everywhere, and stays on your device.

PRIVATE BY DEFAULT
Your conversations and memory are encrypted on your device. Anuma cannot read your data. Open source and closed source model memories are kept separate. Your conversations are never used to train AI models.

COUNCIL MODE
Council Mode sends your prompt to up to four AI models at the same time. 
You get two ways to use it:
Compare: See every model's answer side by side. Claude might give you a nuanced breakdown. GPT might give you a structured list. Gemini might surface something the others missed. You pick the best one.
Unified Answer: Get one combined response that pulls the strongest parts from each model into a  single answer. No copy-pasting between tabs. Use it for important decisions, research questions, anything where a second opinion makes the answer better. Except now you get four opinions at once.

CREATE WITH ANUMA
Generate images, video, and audio with Anuma in one place. Use multiple leading models, keep everything in one library, and create faster with a consistent workflow. Use it for presentations, social media, creative projects, voiceovers, and more.

ANUMA HOUSING AGENT
Anuma Housing Agent helps renters solve housing problems with AI-powered support. Review leases, spot hidden fees, generate demand letters, and get step-by-step help for landlord disputes, rent increases, and other rental issues. Built to make housing help simpler, faster, and easier to understand.

OPEN-SOURCE & CLOSED-SOURCE MODELS
Open source large language models (LLMs) are built on publicly available code and model architectures. On Anuma, open source models run with zero data retention, giving users greater control over privacy and data handling. Closed source LLMs are developed by leading AI companies such as OpenAI, Anthropic, Google, and xAI. These models deliver some of the most advanced reasoning, writing, coding, and analytical capabilities. Depending on the provider and account settings, data sent to closed source models may be retained on their servers and used to improve future models or services.

MORE PRIVATE. MORE PRODUCTIVE. MORE INTELLIGENT.
Join thousands of users who have reclaimed control of their memory and privacy without sacrificing the power of cutting edge AI. With Anuma, your knowledge compounds over time while your data stays yours.

TRY ANUMA FREE
One workspace. One private memory. One bill. All under your control.
FREE credits to try every feature. No credit card required.

Anuma is for anyone who uses AI regularly and wants one place that remembers them, works with every model, and keeps their data private.

Terms of Use (EULA): https://www.apple.com/legal/internet-services/itunes/dev/stdeula/
Anuma Privacy Policy: https://www.anuma.ai/privacy-policy

## Key features

- Access multiple AI models
- Persistent memory across models
- Private and encrypted conversations
- Council Mode for comparative answers
- Unified answer generation
- Image, video, and audio creation
- AI-powered housing agent

## Pricing and in-app purchases

Base price: Free.

**Subscriptions**

| Subscription | Price | Period |
| --- | --- | --- |
| Pro Annual | $204.99 | per year |
| Starter Monthly | $9.99 | per month |
| Starter Annual | $95.99 | per year |
| Pro Monthly | $19.99 | per month |
| Pro Annual | $191.99 | per year |

**One-time purchases**

| Purchase | Price | Type |
| --- | --- | --- |
| 200 Credits | $2.99 | one-time |
| 400 Credits | $4.99 | one-time |
| 1000 Credits | $10.00 | one-time |
| 1,500 Credits | $14.99 | one-time |
| 2,200 Credits | $19.99 | one-time |


## Recent user reviews (1 of 1)

All indexed reviews: https://appshunter.io/ios/app/anuma-private-ai-with-memory/id6758104125/reviews.md

### 2/5 — Too many glitches

*2026-07-05*

I’ll be honest it is nice having Les’s guardrails and having memory merge from different models however this app has way way too many glitches I have to open and close the app every time I do a new message otherwise it just deletes the message after generating also the guardrails are slower getting worse just like the other single models but still a long way to go before it’s guardrail he’ll

**Developer response:** Thanks for the feedback. Some parts of your review are a bit hard for us to fully understand, but it sounds like you may be seeing an issue where messages disappear after a response is generated, requiring you to close and reopen the app before sending a new message. That is not expected behavior, and we’d like to investigate it directly. Please update to the latest version when available, then contact us through in-app support or at support@anuma.ai with your device model, iOS version, account email, and any steps that reproduce the issue. We appreciate you flagging this.

## Frequently asked questions about Anuma

### What AI models can I use with Anuma?

Anuma supports a wide range of leading AI models including ChatGPT, Claude, Gemini, Grok, Qwen, Ling, Kimi, MiniMax, Flux, and Nano Banana. This allows you to leverage the strengths of different models for various tasks without switching applications.

### How does Anuma ensure privacy?

Anuma keeps your conversations and memory encrypted on your device, meaning Anuma cannot access your data. Your data is never used to train AI models, and open-source and closed-source model memories are kept separate for enhanced user control.

### What is Council Mode in Anuma?

Council Mode allows you to send a prompt to up to four AI models simultaneously. You can either compare their answers side-by-side or receive a single, unified response that combines the strongest elements from each model's output.

### Can Anuma remember my preferences across different AI models?

Yes, Anuma Memory stores your preferences and context, carrying them across every model, device, and session. This means your AI already knows who you are and what you're working on, even when switching between models or returning later.

### Does Anuma have ads?

No, Anuma does not contain advertisements. The application is free to use with optional in-app purchases for enhanced features and subscription plans.

### What are the subscription options for Anuma?

Anuma offers several subscription plans, including 'Starter Monthly' at $9.99/month and 'Pro Monthly' at $19.99/month. Annual plans are also available, such as 'Starter Annual' for $95.99/year and 'Pro Annual' for $191.99/year or $204.99/year.

### Is Anuma available on multiple platforms?

Anuma is currently supported on iPhones and iPads, indicated by its availability on these devices. The app is designed to maintain your AI memory across all supported platforms.

### How often is Anuma updated?

The latest version of Anuma is 1.1.2, and it was last updated on June 24, 2026. While specific update frequency isn't detailed, this indicates active development and regular improvements.

## Version history (last 5 releases)

### 1.4.0 — 2026-07-28

Bug fixes and small improvements.

### 1.3.0 — 2026-07-21

Memories topics redesign, social sharing polish, improved error messages, and bug fixes across mobile and web.

- Try the redesigned Memories screen on mobile: topics replace folders, with a floating search dock and a new topic-management flow
- Share generated images and audio to public links on mobile and see the link copied to your clipboard; edit the action row to remove a share
- On mobile, generated images are now named from their caption (e.g. 'a-cuddly-polar-bear-plushie.png') instead of 'generated-image.png'
- Check that council presets on mobile now populate all 4 models for Free/Basic plans instead of 3
- On mobile, the iOS paywall now surfaces loading states and retry options when RevenueCat offerings fail to load
- Verify that error messages for provider/tool/model failures (council mode) are now specific instead of generic, with accurate copy for common cases
- On web, non-image shares (audio, video, apps) now unfurl with a branded per-artifact preview card instead of a generic image
- Mobile now attaches W3C trace-context headers (traceparent) to requests to your backends for distributed tracing
- On web, memory-graph nodes now display distinct icons for organizations, products, and events (matching the expanded entity taxonomy)
- Check that imported memories on mobile and web are filtered to remove junk entries (degenerate fragments, bare names) and empty imports now show 'no memories found' instead of proceeding to review
- Topic extraction now runs continuously in the background on mobile (parity with web), filling topics for manual/imported memories so they're managed by LLM instead of regex

### 1.2.3 — 2026-07-18

• Better Council summaries
• Better memory organization with Topics
• Generate PDFs in chat
• Share AI generated images with a public link
• Personalized prompts for audio and video creation
• New models: GPT-5.6, Grok 4.5, GPT Image 2, and Grok Imagine

### 1.2.2 — 2026-07-14

Try out these improvements in this beta build.

- Scroll through unlimited Recent chats on the home screen — the list now loads 15 at a time and grows as you scroll instead of showing just 3.
- Try the in-app Net Promoter Score survey: you'll see a quick 0–10 rating card on your fifth message in a conversation, then optionally again on Day 2 if you return to the app.
- Search your Memory Vault by meaning, not just keywords — type a query and it ranks matches by semantic similarity to what you're looking for.
- Real attachment previews in the chat: see PDF first pages, video frames, and text snippets inline instead of generic icons.
- Tap a PDF attachment to open a full-screen viewer — pinch to zoom, swipe to flip pages, and swipe down to close.
- Sign in faster — the "Signing you in" loading screen is gone; you stay on the current screen with inline spinners while auth completes in the background.
- Android predictive back gesture now supported — swipe from the left edge to go back, with a visual preview of where you'll land.
- Android: fixed a launch crash caused by a stale dependency version.
- Fixed chat scrolling: the pin spacer below messages now releases once a reply settles, so you can scroll naturally without a blank region in your way.
- Fixed slow image uploads — the app now encodes images to base64 and uploads them in the background as soon as you attach them, so they send much faster.
- Fixed memory recall for Auto model: if you chose "all models" in privacy settings, Auto chats now correctly recall private memories instead of losing them.
- Fixed memory import: memories that fail field encryption are now securely skipped and reported to you instead of silently lost.
- Fixed Slack connection: existing users who granted Slack read permissions can still use Slack tools without reconnecting, even though we now ask for write access on new connects.

### 1.2.1 — 2026-07-08

• Council redesign: ask several models at once and unify their answers
• Privacy toggle to redact personal details before they leave your device.
• New models: Claude Opus 4.8, Sonnet 5, Fable 5, and GLM 5.2.
• Faster streaming, plus fixes for scrolling, navigation, and video generation

## Related topics

[jajaya ai guardian](https://appshunter.io/ios/topics/jajaya-ai-guardian) · [core memories ai](https://appshunter.io/ios/topics/core-memories-ai) · [cverse ai](https://appshunter.io/ios/topics/cverse-ai) · [meance ai](https://appshunter.io/ios/topics/meance-ai) · [renew ai](https://appshunter.io/ios/topics/renew-ai) · [memory aid ai](https://appshunter.io/ios/topics/memory-aid-ai) · [manys](https://appshunter.io/ios/topics/manys) · [fairprice ai mobile](https://appshunter.io/ios/topics/fairprice-ai-mobile) · [jublia ai invietnity](https://appshunter.io/ios/topics/jublia-ai-invietnity) · [frugalfox ai](https://appshunter.io/ios/topics/frugalfox-ai) · [sparrow ai](https://appshunter.io/ios/topics/sparrow-ai) · [reroom ai free](https://appshunter.io/ios/topics/reroom-ai-free) · [unmix ai](https://appshunter.io/ios/topics/unmix-ai) · [dapta ai](https://appshunter.io/ios/topics/dapta-ai) · [super grok ai](https://appshunter.io/ios/topics/super-grok-ai) · [breakfree ai](https://appshunter.io/ios/topics/breakfree-ai) · [canva ai free](https://appshunter.io/ios/topics/canva-ai-free) · [notable ai](https://appshunter.io/ios/topics/notable-ai)

---

*Data collected daily from the US App Store and indexed by [AppsHunter](https://appshunter.io/). User reviews are verbatim App Store reviews. Ratings, prices and chart positions refresh continuously; this snapshot is from 2026-07-31.*