# Mollama — user reviews > 50 indexed App Store user reviews of Mollama by 飞 杨, rated 3.60/5 from 29 ratings. - Source: https://appshunter.io/ios/app/mollama/id6736948278/reviews - App overview in markdown: https://appshunter.io/ios/app/mollama/id6736948278.md - Rating breakdown: 5★ 13 · 4★ 4 · 3★ 6 · 2★ 0 · 1★ 6 ## What do users think of Mollama? Overall sentiment: negative **Pros** - 👍 Great for local models - 👍 Native MLX support - 👍 Simple interface - 👍 Good for hobbyists - 👍 Works with Llama - 👍 OCR functionality - 👍 Offline LLM capability - 👍 Good for small tasks - 👍 Private AI assistant **Cons** - 🐛 Crashes frequently - 🐛 Buggy and error-prone - 🐛 Unusable with larger models - 🐛 Download issues - 🐛 Slow downloads - 🐛 Inaccurate download progress - 🐛 Models not loading - 🐛 Unsupported model types - 🐛 App freezes phone - 🐛 Keyboard overlaps chat - 🐛 Can't load external servers - 🐛 Doesn't work offline sometimes - 🐛 Performance lag - 🐛 App quits unexpectedly - 🐛 NSURLErrorDomain errors - 🐛 Limited model selection - 🐛 Chat window issues - 🐛 Doesn't send prompts when long - 🐛 Lays MacBook Pro low **Commonly reported bugs** - App crashes on sending messages with certain models - Tokenizer metadata errors in offline mode - Unsupported model type errors (e.g., Gemma, Qwen3) - Models not loading when selected - App freezes phone during model download - Download bar lags app significantly - Downloads stop when app is inactive - Downloaded models stop working offline - App crashes when trying to run chain-of-thought models - App crashes with Qwen2.3 35b - Keyboard overlaps chat box and send button - Cannot load external servers - App crashes when trying to chat - App crashes on newer generation iPad with Qwen2.3 35b - App crashes on older iPhone 11 Plus - Ollama access not working - App crashes when trying to send prompts in long chats - Camera button in chat window doesn't work **Most requested features** - Add notification permission for background downloads - Show LLM name for each answer - Support for Qwen3 family of models - Ability to use any HuggingFace MLX models - Allow user-defined model folders - Adjustable model parameters (temperature, etc.) - Shortcut support - Support for interacting with shortcuts - Ability to modify system prompt - Add a donate button ## Newest reviews (14 of 50) ### 4/5 — 请问下载模型一直没进度怎么处理 *2025-09-19, version 1.2.2* 请问下载模型一直没进度怎么处理，科学上网也不行，没有进度，一直下载不成功，求解决办法 ### 3/5 — Needs Search Filtrting *2025-09-18, version 1.2.2* An amazing App that allows you to download any model really, which deserves high praise. But *any* model includes ones that in no way will fit or run on your phone. ### 3/5 — Turbo unusable *2025-08-13, version 1.2.2* Ollama turbo isn’t usable currently it seems like. Gives a NSURLErrorDomain ### 3/5 — Full of error *2025-07-27, version 1.2.2* Not Good. ### 3/5 — Has potential, but is unfortunately buggy… *2025-06-12, version 1.2.1* Downloads of MLX models in the app is extremely slow (even with high-speed internet), download progress isn’t accurate (it may still show download in progress, when the model is already downloaded), and once models are downloaded, they aren’t loaded when selected (system and app memory usage remains unaffected when selecting larger models). When a non-built-in model is selected, and the simple prompt "hi" is given, the app is busy, and returns nothing, except at some point flashing an orange "unsupported model" (or something like that" message. Things work with the built-in model, but I don’t need an MLX app to run something as trivial as a 0.5B micro model on an M4Max with 128GB RAM, when I could run that model on an old iPhone… If only the bugs were fixed, and this would be a great start. ### 4/5 — Great for everything except Gemma *2025-06-05, version 1.2.1* This app worked great for every model I tried, until I tried a small Gemma model. Gemma is literally unsupported (gives error "type gemma-text is not supported" or similar) and it's one of my favorite local models. It runs fine on my much older Android phone so I don't understand why Gemma is unsupported. Other than that, this is my favorite native AI app on iOS. ### 1/5 — Très déçu *2025-05-14, version 1.2.1* Téléchargement monte par 0,01 % à la fois unemfois tous les 10 secondes, et le téléchargement plante très tôt, dommage, il à supposément accès à plusieurs LLM ! ### 4/5 — Great offline app *2025-05-09, version 1.2.0* It would be perfect if the app shows the name of LLM for each answer, since there are many models to choose from. ### 3/5 — Newer models unsupported *2025-04-28, version 1.2.0* Newer models such as qwen 3 give error message “Unsupported model type: qwen3” whenever you try to send a chat to them ### 4/5 — Great app for mlx open model testing *2025-04-27, version 1.2.0* This app is what I was looking for my MLX LLM models. Just needs the ability to use any huggingface MLX models for any mlx repo url , and a model folder user can use their own trained mlx models .Also would be great to be able to adjust model parameters eg temperature etc . Thanks for your great work ### 1/5 — Hängt *2025-04-14, version 1.2.0* Die app hat mein MacBook Pro M3 Max lam gelegt. ### 1/5 — Crashes *2025-04-06, version 1.2.0* iPhone 16+…. I downloaded six or more models and all of them crash doesn’t work even once as small as 1b ### 3/5 — Needs shortcut support *2025-04-03, version 1.2.0* Runs MLX using native Apple hardware but lacks shortcut support ### 2/5 — Funktioniert leider nicht. *2025-03-22, version 1.2.0* Auf einem alten iPhone 11 Plus funktioniert die Locke MLX Funktion nicht, nachvollziehbar, man könnte allerdings informiert werden, kein Grund die App abstürzen zu lassen. Das sollte abgegangen werden. Leider funktioniert der Zugriff auf OLLAMA nicht welches ich auf meinen MAC hoste nicht. Weder lokal im Heimnetz noch über ngrok veröffentlicht. Die MAC Funktion funktioniert mit localhost. ## Most critical reviews (10 lowest-rated) ### 1/5 — As useful as a notepad *2026-03-21, version 1.2.2* Didn’t do anything, couldn’t figure out anything ### 1/5 — Move on…there are way better apps *2026-02-25, version 1.2.2* Buggy and error-prone on both macOS and iOS. Seriously, there are a dozen better apps. Save your time. ### 1/5 — Terrible, terrible terrible app *2025-11-10, version 1.2.2* Virtually every question I pose to the app came back with the same response: I can’t answer that question. When queried further, it would not tell me why. ### 1/5 — Can’t download *2025-09-04, version 1.2.2* Can’t download more than 90.91%…all model same ### 1/5 — Cannot do one thing *2025-09-04, version 1.2.2* Downloaded it, asked it 4 different questions, eg summarise this document etc. Just continually crashes and will not function at all. Very disappointed. ### 1/5 — Doesn’t work *2025-08-29, version 1.2.2* This app doesn’t work at all. It just quits unexpectedly whenever you enter a query. No error message, no nothing. ### 1/5 — Extremely buggy and basically unusable *2025-07-29, version 1.2.2* I have never written an app review before but this app has made me so mad that I have decided to change that. This app is in a terrible buggy state that is awful to use even with small models! On an iPhone 16 pro max (8gb of ram) the built in model (around 200mb) will lag when generating a long response and the app will eventually crash. This happened no matter the model or how powerful your device is. Only thing worse than that is downloading new models. The DOWNLOADING BAR lags the app to 5 fps when in view and has crashed several times for me. You also cannot exit the app or even the loading view or the download will just stop. The loading bar is also inaccurate with models finishing downloading before the bar hits 10%. Downloading in general takes forever and it will not tell you if you can even run a model until it is fully downloaded. Also out of no where downloaded models will stop working offline with an error saying it can’t fetch some txt file for the model which ruins the point of the whole app. Do not use this app if you value your sanity. ### 1/5 — Системная ошибка *2025-07-03, version 1.2.2* Не работает , вылетает при первом запросе ### 1/5 — Nope *2025-06-06, version 1.2.1* Does not work, can’t load any external servers, can’t display models, it connects, I see the interaction, the server is sending the models but app can’t see them at all ### 1/5 — Crashes Instantly with Local Models – Unusable *2025-05-22, version 1.2.1* Tried loading the “Mistral-7B-Instruct-v0.2-4bit” model, and the app crashes immediately upon sending a message. Even offline mode throws tokenizer metadata errors. No clear documentation, no fallback behavior, and no error handling. Completely non-functional for intended use. Needs urgent stability fixes and proper tokenizer integration. One star until it actually runs a model without crashing. ## Most praised reviews (10 highest-rated) ### 5/5 — Awesome *2025-09-26, version 1.2.2* Love being able to run local models via MLX. Please add a notification permission request to allow downloads to continue in the background ### 5/5 — 好用 *2025-06-05, version 1.2.1* 这是我用过最好用的，既能跑本地又能跑api的私人人工智能助手应用，界面也很简洁，美中不足就是还不支持全模态，感谢作者！ ### 5/5 — Great app! *2025-05-21, version 1.2.1* This is a great app! Thanks for making this. ### 5/5 — Really neat on a phone. 5 stars for native MLX *2025-05-10, version 1.2.1* This is the kind of the thing that’s intended for people who mess around with these things as a hobby, but it works really well as long as you don’t try to use models too big to fit in your phone’s RAM. You’re probably gonna be limited to sub-4B models on an iPhone, but it works really well for what it is. And for what it’s worth, it does have a very real world use for me because sometimes I don’t have good access to data and can’t just use the GPT app. I hope the developer continues to work on this. If they add a donate button I’ll buy them a coffee. ### 5/5 — Qwen3 *2025-05-01, version 1.2.0* Please add support for the new Qwen3 family of models! ### 5/5 — Best app in ios *2025-04-23, version 1.2.0* The best is the simple. Like this app. Would participate or donate if it is open source. ### 5/5 — Une bonne petite AI *2025-03-04, version 1.1.2* une bonne petite AI pour les petites taches discute rapidement et aide à trouver des solutions. Je recommande. Merci au développeur ### 5/5 — Works good so far *2025-02-04, version 1.1.2* I’m using it for a week now and it works just fine. The OCR functionality is quite cool. ### 5/5 — Offline LLM on your phone *2025-02-04, version 1.1.2* This app works amazing. I was not expecting the LLM to work this well on my iPhones hardware ### 5/5 — Excellent *2025-02-02* Works with llama. I have downloaded mlx file via LLM studio, but Mollama is unabe is unable to search or locate file. Is it possible to point app to mlx file? Update: Tested with MLX - straight forward - many thanks for free app --- *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 2026-08-11.*