Great
Amazing
Amazing
Incredibly powerful app. It truly hits different to talk to an AI running on your smartphone when you’re out in the middle of nowhere with no internet connection whatsoever. The models are very impressive with their capabilities for the small parameter count.
REQUEST: please allow us to switch the default model. It resets after every new chat! When using this with a custom backend it takes like 4 clicks each chat just to switch it, which makes an otherwise great feature more annoying to use.
REQUEST: please allow us to switch the default model. It resets after every new chat! When using this with a custom backend it takes like 4 clicks each chat just to switch it, which makes an otherwise great feature more annoying to use.
Show more
Brilliant Concept (I have some suggestions)
Honestly, this app utilizes “Apple Intelligence” (or, Apple’s integrated support for local AI via certain hardware and software additions and changes) far better than any other application. It literally allows you to use supported models locally, on device, free of charge. Apple themselves is yet to even improve Siri and allow their own very flawed voice assistant to make use of this. I paid for the new iPhone assuming this would be utilized, and so far, this app is probably the only application which makes ample use of that.
Suggestions:
- Enable sideloading of supported models by means of HuggingFace access (gguf, safetensor, etc… whatever is supported)
- Addition of more support for inference or API based model providers (aside from OpenRouter, perhaps Kimi/Z.AI/OpenCode’s plans which offer more reasonably priced access, Venice AI, etc)
- Addition of more tools (such as how the web search tool works; adding more would be nice)
- Enable MCP support (the search tool is cool, but not good enough — enabling the use of MCPs could help; perhaps add an addon store)
- Enable scripting for custom agents, tooling, etc (I’d prefer an “OpenCode”-esq agentic experience over the current agent, so customization of behavior, guiding reasoning, allowing custom tooling using premade SDK and otherwise, etc would be very cool)
I was thinking about making an open source version of this app and integrating such features, but clearly the developers of this application are competent and can do it themselves as they wish.
Suggestions:
- Enable sideloading of supported models by means of HuggingFace access (gguf, safetensor, etc… whatever is supported)
- Addition of more support for inference or API based model providers (aside from OpenRouter, perhaps Kimi/Z.AI/OpenCode’s plans which offer more reasonably priced access, Venice AI, etc)
- Addition of more tools (such as how the web search tool works; adding more would be nice)
- Enable MCP support (the search tool is cool, but not good enough — enabling the use of MCPs could help; perhaps add an addon store)
- Enable scripting for custom agents, tooling, etc (I’d prefer an “OpenCode”-esq agentic experience over the current agent, so customization of behavior, guiding reasoning, allowing custom tooling using premade SDK and otherwise, etc would be very cool)
I was thinking about making an open source version of this app and integrating such features, but clearly the developers of this application are competent and can do it themselves as they wish.
Show more
Bonsai teraney 8B
Please add bonsai teraney llm 8B as the only app that has it is local ai by lm studio which its system prompt only allows 1000 tokens I use very advanced system prompts and need infinite system prompt token limit. Please add this model as its fast and works well on my iPhone 13 Pro
Either adopt AI or don’t; this is not a viable alternative
I’ve tried about every AI tool imaginable over the past 2 years and based on my experience, I am dumbfounded as to why this is so highly rated and now with Apple Intelligence becoming at least functional over the past few weeks, why this will remain an option in the iOS store. There’s legitimately not one use case for it since it botches everything you ask it to do. That said, I also understand the intention behind the app and applaud that but by AI’s very nature, it’s efficacy is inversely correlated to privacy —the more privacy protections you’re willing to waive, the better the underlying AI will perform; nothing will change that as Edge AI is largely a farce
@Apple - stop censoring my reviews. The next time you do I’m reporting your conduct to the AG’s and FTC’s offices
@Apple - stop censoring my reviews. The next time you do I’m reporting your conduct to the AG’s and FTC’s offices
Show more
World Changing
Liquid’s models are going to change the world. Imagine the entire power of Anthropic plus Open AI in the palm of your hand powered only by your phone.
Good, but…
It’s a good app, but what is going on with the text. It’s so small and just overall hard to read chat interface. If they can fix that it’s 5 star. Even if you adjust the size it gets all out of place when you try and select text. No idea what is going on, does nobody test it haha
Could have more
Could have a lot more features, like being able to download your own custom model or import your own gguf or safetensors file. It’s a good nice designed app though i would like more customization options for it
Does not work
I paste my OpenRouter API key and select a model. Any model. The result is always the same, a vague error that there is a provider error and no clue what that error is. No help, and contact is only through a small form. Some help would be nice, or at least a clue on what the error is.
Janky
Web search doesn’t work most of the time and image generation doesn’t work at all.
Solid
I really like it. But I really need a "thinking effort" setting.
Minimalistic and powerful
Everything perfect. The only thing i miss is the setting to hide reasoning process.
Great idea, needs to mature a bit
Still some quirks to work out in the UI, inconsistencies, weird UX flows, etc. overall, love that you can download models to run locally (even though it's only LFM and Qwen 3 models), and can connect to an openai compatible custom endpoint. Even comes with web search built in. 5 stars if the custom endpoints also accept images as input.
Show more
Bad Context size
I have tried to add some promts for better conversation, but any bot starts to not see my messages, he can only see promt. The context size is very small, and also I think that instructions are sending to chat every time I try to message my questions, not only 1 time at the start of conversation. Better use ChatBox
Show more
crashes
The app crashes on my iPhone 16 Pro Max when I use a custom backend the answer comes back but immediately after that crashes, it seems to work fine on my iPad Pro M2
Works very well, much better than other similar apps such as LocallyAI
Title says it all.
My new Goto
I feel like i have tried every local AI app in the app store and this one punches beyond anything i have tried. On iPhone 14 pro, so outdated and yet the experience is flawless. Web search works better than most I’ve used. All the models are quick and optimized. I am looking forward to the addition of search for vision capabilities. Once that is implemented in both vision and voice chat, we have a winner.
Show more
Can’t retain temp settings
As title says
More models
I know this is made by Liquid AI and they want to promote their model, but this is the best mobile UI for local AI. Since you already have qwen, please also add gemma 4 if possible.
good job. need a terminal plugin
這正是所需的：能夠同時運行雲端和本地大模型。如果能有直接運行大模型的終端插件或腳本，那就更好了。
This is exactly what is needed: the ability to run both cloud and local LLMs simultaneously. It would be even better if there were a terminal plugin to run the models script directly.
This is exactly what is needed: the ability to run both cloud and local LLMs simultaneously. It would be even better if there were a terminal plugin to run the models script directly.
