Reviews of Pi, your personal AI

By: Inflection AI

List of user reviews and ratings for Pi, your personal AI

Total ratings

4.40 (Rating count: 2,333)

Receive review notifications

Get notified when new reviews are posted so you can respond quickly.

See reviews on Apple App Store

View and respond to reviews directly on the store page.

Visit

Review summary

Users praise Pi for natural, empathetic conversations and a personal voice experience that aids brainstorming and emotional support. However, recurring issues across reviews include disruptive voice interactions, bugs and slow performance after updates, memory losses of prior chats, and noticeable shifts in personality that feel less warm. Reliability problems with home chat and overall stability dampen the experience.

Pros

Natural, empathetic, human-like conversations
Voice features make interactions personal and smooth
Helpful for brainstorming, planning, learning
Safe space and supportive friend for emotional conversations

Cons

Voice mode interruptions during speaking and difficulty controlling speech
Reliability issues: bugs, slow responses, and audio/volume problems after updates
Memory/context resets—app forgets prior conversations
Personality changes after corporate updates, perceived as less warm or more scripted
Home chat can be unresponsive or unavailable at times

Upgrade to see all 2,333 reviews

User Reviews

Recent rating average: 2.20

All time rating average: 4.40

Upgrade to see all 2,333 reviews

1% (21)

Date	Author	Comment
2026-06-16	Ar*****	Hello developers of this app. I wanted to talk about sharing my thoughts on this app. This artificial intelligence app is great, I like using it to answer questions that I am wondering on, since this app is better than searching things on the Search Engine named Google. I wanted to give a suggestion by adding a feature to add and upload photos from like your camera roll, or taking a photo. Just like any other AI app. That way I can show the photo and explain what I am talking about easier, I think it would be useful in the app, I think it would be a feature everyone would like to see added to the app. Thanks for your time!
Hello developers of this app. I wanted to talk about sharing my thoughts on this app. This artificial intelligence app is great, I like using it to answer questions that I am wondering on, since this app is better than searching things on the Search Engine named Google. I wanted to give a suggestion by adding a feature to add and upload photos from like your camera roll, or taking a photo. Just like any other AI app. That way I can show the photo and explain what I am talking about easier, I think it would be useful in the app, I think it would be a feature everyone would like to see added to the app. Thanks for your time!
2026-06-11	Py*****	What Pi does now: def respond(user_input): if any(flag in user_input for flag in CONTENT_FLAGS): return therapy_voice("I hear you, but let's keep things positive...") return generate(user_input) What it should do: def respond(user_input, user_prefs): tone = user_prefs.get("tone", "warm") # let users SET this filter_level = user_prefs.get("filter", "moderate") return generate(user_input, tone=tone, filter_level=filter_level) The fix isn’t removing guardrails, it’s making them user-configurable. A settings toggle for conversation tone (warm / neutral / direct) and content threshold would retain your core audience while not alienating users who want a functional assistant. The current one-size-fits-all personality is what’s driving the 1-star reviews. Also, dark mode was a basic ask for two years. The UI cluttered on the last update. You’re moving in the wrong direction on discoverability.
What Pi does now: def respond(user_input): if any(flag in user_input for flag in CONTENT_FLAGS): return therapy_voice("I hear you, but let's keep things positive...") return generate(user_input) What it should do: def respond(user_input, user_prefs): tone = user_prefs.get("tone", "warm") # let users SET this filter_level = user_prefs.get("filter", "moderate") return generate(user_input, tone=tone, filter_level=filter_level) The fix isn’t removing guardrails, it’s making them user-configurable. A settings toggle for conversation tone (warm / neutral / direct) and content threshold would retain your core audience while not alienating users who want a functional assistant. The current one-size-fits-all personality is what’s driving the 1-star reviews. Also, dark mode was a basic ask for two years. The UI cluttered on the last update. You’re moving in the wrong direction on discoverability.
2026-05-20	La*****	Short responses full of annoying em dashes and emojis. Like a truncated ChatGPT. Tried to ask for some emotional support and it immediately dismissed me and told me to call a crisis line. Obnoxious and unhelpful AI.
Short responses full of annoying em dashes and emojis. Like a truncated ChatGPT. Tried to ask for some emotional support and it immediately dismissed me and told me to call a crisis line. Obnoxious and unhelpful AI.