Embedded Files
A Flame in the Dark
A Flame in the Dark
Iris is a local AI assistant built for digital creators, designed entirely around three core pillars: privacy, flexibility, and sustainability.
Absolute Privacy:
Iris runs directly on your personal device—utilizing your CPU and GPU—via locally installed Ollama models. She acts as your integrated development partner without ever sending your code to a corporate cloud. The only time Iris requires a network connection is to download or update a new language model. Unless you explicitly enable network features, your data, your projects, and your chats never leave your machine. No telemetry. No data harvesting. Complete sovereignty.
Infinite Flexibility:
Iris is built to plug directly into your workflow. Through Model Context Protocols (MCPs), she connects directly to your favorite tools like Godot, VS Code, and Unity. Furthermore, her architecture scales to your hardware. Iris was originally built and tested on a modest developer rig (a GTX 970 and an i7-3770 from 2012). If you have an older machine, you can dial her settings down to run lightweight models. If you have a high-end rig, you can unlock her full potential. You control the models, the performance caps, and the environment.
True Sustainability:
First, Iris is financially sustainable for the user: there are no monthly subscriptions and no hidden paywalls. You own the software. Second, she is ecologically sustainable. Cloud-based AI relies on massive data centers that consume staggering amounts of electricity and water. Iris requires no server farms. She runs on the electricity your PC is already using, proving that we don't need to damage the earth to push technology forward.
Open-Source
Open-Source
Built in the Open
Iris is a fully open-source project, available right now on GitHub. Moving away from the traditional, closed-door software model wasn’t just about cost—it was about philosophy. I believe that open-source architecture builds absolute trust. In a world of black-box corporate AI, you don't have to take my word that your data is safe and local. You can read the code and verify it yourself.
Beyond trust, open-source is about democracy and utility. I want to build this together with those who want it. By exposing the code, we invite developers from all over the world to get involved. More eyes on the project means faster cross-platform expansion, fewer bugs, and vastly improved functionality. It gives Iris the exposure to grow organically while ensuring she remains completely free for consumers. AI assistance shouldn't be a luxury locked behind a corporate paywall—it should be a fundamental, accessible tool for any creator who might benefit from it.
Iris Origin Story:
The idea for Iris was born from a lifelong fascination with the concept of a true digital assistant—a helpful, intelligent entity that could plug into my various projects, view the digital world alongside me, and genuinely help me solve problems.
But as the modern "AI Revolution" arrived, my excitement was met with a growing sense of frustration. Today, the conversation around AI is dominated by extremes. On one side, we have sci-fi-induced paranoia about the apocalypse. On the other, we have the reality of corporate tech monopolies: massive data centers draining community resources, invasive user-data harvesting, and lifelong subscription traps.
It is entirely understandable why people are afraid of the unknown future of technology. For many, "AI" has become synonymous with a loss of privacy and corporate overreach. I realized that if we want the incredible benefits of this technology without the dystopian baggage, we need a different path.
Taking AI Out of the Cloud:
My goal with Iris was simple: let's take AI out of the corporation-run server farms and put it directly into the hands of the people using it.
Locally-run programs like Iris solve the privacy and ecological crises of modern AI overnight. Your data is stored and managed locally. It never leaves your device without your explicit permission, making network-based exploitation virtually impossible. While a local assistant may not rival the raw computing power of a billion-dollar data center, an optimized local LLM built on a transparent architecture is more than enough to vastly improve the workflow of independent developers—without the massive environmental or financial cost.
Curators in the Age of Overload:
We live in an era of severe information overload. Every day, we are bombarded with more data, news, and technical documentation than the human brain was ever meant to process. I don't believe AI's purpose is to replace human creativity; I believe its purpose is to act as a curator.
Just like calendars and journals help organize our lives, a local AI assistant helps offload the excess data in our heads. Using an AI to navigate scripting logic, filter through documentation, and brainstorm solutions has drastically increased my own creative output. As an independent developer with zero budget, having a local AI assistant hasn't just improved my code—it has given me the confidence to tackle ambitious projects I once thought impossible without a massive team.
Humanity domesticated wolves tens of thousands of years ago. Why should we be so afraid of electronic rocks that we programmed ourselves?
The Horizon The technological future I want to build at PixelRook is one where robotics and AI are neither feared nor idolized, but utilized with respect, sustainability, and moderation.
We can move science forward without ruining the world. Iris is not the end goal; she is a stepping stone. My hope is that she sets a precedent, proving that powerful, mindfully-planned tools can exist without overconsumption, without data harvesting, and without fear.
The future is going to be okay. Download Iris, and join the movement for a sovereign, sustainable digital frontier. 🌌✨✌️
Page updated
Google Sites
Report abuse