Your Personal Private & Local AI Companion. Chat with open-source models on device, connect to online AI services or private LLM servers. Dev: @stevenmoon
Major PocketMind update is live!
Adds MLX support, random prompt generator and much more.
PocketMind is free to download.
Get your own private and local AI now.
Currently there are 7 new apps that are on testflight beside the completely rewritten version 2 of the OG PocketMind my two new favorites are:
https://t.co/W4HS8DzQen
https://t.co/LQ4GroSuQH
Created an open source ai chat bot.
- you can save and edit data locally in the browser
- you can add your openai, google gemini or antrophic api keys and the AI chat app is functional running it locally
- has chat session history
- light and dark mode
- mobile optimized
You can run it locally in a browser at the demo website or download and run the code from git hub.
Adding localization to @pocketmindapp using Gen AI for Spanish, Chinese, German, Japanese, French, Arabic, Russian, and Hindi.
Some of the sample prompts are a bit silly—I’m sure plenty of that will get lost in translation.
What could go wrong? Hopefully nothing too offensive.
Bracing myself for some interesting support emails in languages I don’t understand.
This may not end well. Time will tell.
New TestFlight available for @pocketmindapp!
Your private, local AI chatbot just got better.
• Now supports 8 additional languages (thanks to o3-mini)
• Added a new default model card for o3-mini in the online services section (bring your own API key)
Check it out!
@SauronsGaySon @bonda_soup@AravSrinivas The tip feature just went live today. Based on launching 100s of apps running my app dev agency @clevercoding I don't have much hope for the tip feature.
But if I can bring in enough revenue from tips then I can keep the app free. from: @stevenmoon
@bonda_soup @SauronsGaySon @AravSrinivas No catch. It is free (for now). I plan to add some paid feature down the road but right now just trying to make it the best local and private AI Chat app on iOS.
I'm betting on the fact that I think private and local AI is the future. From: @stevenmoon
The screenshot below shows what can happen when what you are building can be attached to a trending topic.
I release the @pocketmindapp app a couple of months ago and have had slow stead growth.
I have been posting into what feels like a void for a long time on X and LinkedIn.
But last week I added support to run DeepSeek R1 locally and privately on an iOS device.
After the app update hit. My organic search listing in the App Store and my app downloads skyrocketed.
I have been posting and replying like crazy the last two days so this boost involved a lot of work by me.
But is is cool to see the potential of what can happen when your product can incorporate some features that touch on trending topics.
I guess the moral of the story is keep building cool stuff and talking about it. Do all you can to provide value to your users and just keep at it and don't give up.
Thank you to all the new users of the @pocketmindapp app. This is a side project for me and I do it mostly so I can learn how to build AI powered iOS apps.
I appreciate all the feedback I have gotten so far and the many users I have gotten to talk to.
Thanks for your patience as I work through the issues to make this the best iOS app to run local AI models.
A PocketMind update is live.
It added a FAQ and Help section.
The streaming is smoother and scrolls automatically.
The app is free. So I added a "Tip Developer" section with a bunch of in app price options to send me a tip and some crypto addresses people can send a tip to.
Thanks for all the feedback from all the new users I have gotten to talk to. PocketMind is currently a side project that I put a ton of time into. I am just one person so I appreciate the patience users have had with me and I work on making this the best option to run AI models locally and privately on your iPhone.
@GeorgeSiosi@stevenmoon@Benioff@lmstudio We don't have an android app yet. If we can get enough traction with the iOS version we will look at bringing the app to Android.
The app can run models using the llama.cpp engine or the MLX engine.
https://t.co/sbsLWahUJ6
https://t.co/eveUl95wjI
Then all the models are off hugging face.
For example MLX has a great community on hugging face which is where most the mlx models come from.
https://t.co/jw1F7wwiQp