🌍 Environmental Impact

Every time you use ChatGPT, Gemini, or Claude, a request travels to a massive datacenter and comes back. Those datacenters run thousands of GPUs at full power 24/7. The GPUs generate enormous heat. To prevent them from burning out, the facilities pump millions of liters of water through cooling systems every day. That water evaporates. It doesn't come back.

IHYIILOVEYOU AI runs on the device you already own. The heat from your computer or phone is the same heat it would have produced anyway. No extra datacenter. No extra cooling. No water consumed. Every query you run here instead of a cloud service is a real, measurable reduction in water use.

🔒 Privacy You Can Actually Verify

When you type something into ChatGPT, it leaves your device immediately. It goes to OpenAI's servers, gets processed, gets logged, and may be used to train future models depending on your settings. The same is true of every cloud AI.

With IHYIILOVEYOU AI, nothing leaves your device. Your conversations, your learned facts, your stories — all stored locally in a database only you can access. You can delete your account and every byte of your data is gone. Not archived. Not anonymized. Gone.

🚀 Help Us Get Better Hardware

Right now IHYII runs on a 70B model — already powerful enough to impress most people. But the AI world doesn't stop at 70B. Models like Llama 3.1 405B are nearly indistinguishable from GPT-4 in everyday use. Running one of those privately, with zero data leaving your device, requires dedicated hardware that costs real money.

Every donation directly funds better AI for everyone who uses this app — not a corporation, not shareholders. One developer, one mission: AI that doesn't cost the planet its drinking water.

Being Honest: Where Cloud AI Still Wins

Raw capability right now: GPT-4 and Claude are trained on more data with more compute. On the hardest reasoning tasks, they have an edge. IHYII with a 70B model is very capable — most people genuinely can't tell the difference for everyday use — but the gap is real for the most complex analysis.

Real-time web: Cloud AIs can browse the internet live. IHYII uses Wikipedia and its knowledge base. It knows a lot, but it's not pulling today's news in real time.

The gap will narrow: With donations and community support, IHYII will move to dedicated hardware running 405B+ models privately. What runs locally today was completely inaccessible to individuals just a few years ago. The trajectory is clear — and your support is what makes it happen faster.

What are all the tabs for?

Same AI brain, different focus modes. Each tab sets the AI up for a specific kind of task:

Is it the same AI in every tab?

Yes — one brain, different contexts. Think of the tabs like different apps on your phone. Story mode maintains a running narrative history. Questions mode pulls external sources. DevMode writes tighter, more technical answers. Chat is the general-purpose everything mode.

What powers the AI?

IHYIILOVEYOU AI runs on a large 70-billion-parameter language model. It's already connected and ready — just start chatting. Switch between speed and depth using Think Mode in the top bar: lighter modes answer fast, deeper modes think more carefully through harder questions.

How do I teach it things about me?

Just tell it naturally: "My name is Alex", "I prefer dark themes", "I'm working on a novel called The Last Signal". It saves these automatically and uses them in every future response. You can view everything it's learned in the Memory tab.

Is my data private?

Everything is stored on your device or your own server. Nothing goes to Google, OpenAI, or any other third-party cloud. Your conversations, learned facts, and account data never leave the machine running this app.

How do I report an error?

Hover over any AI response and click the 🚩 flag that appears on the right side of the bubble. Choose a category, describe what went wrong, and hit Submit. Reports go directly to the developer's debug panel.

How do I change how the app looks?

Two ways: Themes tab for full color schemes (129+ options), or the 🎨 Customize button in the top bar for layout controls — tab position, chat style, bubble shape, and your own color palette.

What languages are supported?

23 languages in the UI dropdown — English, Spanish, French, Portuguese, German, Italian, Japanese, Korean, Chinese, Russian, Arabic, Hindi, Indonesian, Vietnamese, Swahili, Turkish, Dutch, Polish, Swedish, Bengali, Ukrainian, and Malay. When connected to the 70B model, it can respond fluently in dozens more.

Does it work offline?

The heuristic brain — which handles math, unit conversions, Wikipedia lookups, built-in knowledge, and most everyday questions — works without internet. The 70B language model needs the Ollama server running. The Draw tab is fully offline.

Debug access

Accounts with debug mode unlocked.

All users

Every registered account.

🚩 User Reports

Error reports submitted by users.

🛡️ AI Content Reports

Messages auto-flagged. Click to take action.

📋 Mod Actions