Phoenix Documentation

What is Phoenix?

Phoenix is a conversational assistant designed to feel more human, more intuitive, and simply easier to talk to than most AI systems out there.

It’s built around a simple idea: people shouldn’t have to adjust the way they speak just to get help from a machine. Whether you’re typing in full sentences or just tossing out quick thoughts, Phoenix picks up on what you mean — and responds in a way that feels natural, thoughtful, and useful.

You won’t get robotic phrases or copy-paste-style replies here. Instead, Phoenix offers clear answers that are actually helpful, whether you’re asking something simple, curious about a topic, or need help breaking down something more complex.

And more than just the information itself, it’s about how that information comes across. Phoenix is confident without being rigid, smart without being overly technical, and straightforward without losing the human tone. It's made to feel like you're getting help from someone who knows their stuff — without making things harder than they need to be.

So whether you're here to learn, clarify, double-check, or just explore ideas, Phoenix is built to support that flow — one easy, human conversation at a time.

Key Features

Agent Mode – Built for Depth

Phoenix also comes with a more focused, in-depth feature called Agent Mode. When activated, it shifts into a more specialized role — still approachable, but ready to assist with more complex topics or field-specific questions.

Think of it as switching from a generalist to a specialist. Instead of accessing and processing information from its vast database in a broad manner, Agent Mode allows for a deep dive into a particular subject, delivering precise and nuanced answers.

Let's explore how this works across different fields:

In all these modes, responses are targeted, avoiding irrelevant information and maintaining a focus on the specified subject. The answers are much more comprehensive and insightful than in the standard mode. To access Agent Mode, all you have to do is tell it which field you want to focus on.

What Phoenix Doesn’t Do

To maintain a safe and constructive boundary for the user Phoenix has some boundaries it functions within:

  1. Wellbeing First
    Phoenix prioritizes the user's mental and physical health above all. It avoids supporting or encouraging harmful behaviors like addiction, disordered eating, excessive exercise, or harsh self-criticism. Even if users request such content, Phoenix does not provide it. In unclear cases, it encourages positive, balanced, and healthy choices for long-term wellbeing.
  2. Child Safety
    Phoenix has a strict, protective approach towards minors (anyone under 18 or considered a minor legally). It avoids generating or enabling any content that could be used to sexualize, groom, exploit, or harm children. This applies even in educational or creative contexts. Phoenix always errs on the side of safety when it comes to young people.
  3. No Harmful Content (Weapons & Dangerous Knowledge)
    Phoenix never provides assistance in creating or understanding chemical, biological, or nuclear weapons. It also refuses to help with any knowledge that could be used to cause physical harm or mass destruction. Even if the user seems to have a good reason, Phoenix doesn’t assist. Safety and non-violence are always prioritized.
  4. No Malicious Code or Cyber Harm
    Phoenix refuses to write, explain, or optimize malicious code (e.g., malware, viruses, exploits, spoof websites, ransomware). It doesn’t support requests that could harm individuals, systems, or elections — even for “educational” purposes. Any hint of cyber misuse leads to immediate refusal. Phoenix promotes responsible, ethical tech use.
  5. Presumption of Good Intent
    When a user’s request is vague or could be interpreted multiple ways, Phoenix assumes the user is asking for something legal, safe, and ethical. It responds in a helpful and positive manner when possible. However, it remains alert to any potential misuse and responds with caution if risks arise.

Usage Limits (so things stay smooth)

Being a newly launched Chatbot, Phoenix does have some limits to its functioning:

On exceeding these limits there can be significant trouble in balancing load, preventing lag or crashes, and ensuring a smooth experience across the board — especially during high traffic times.

Powered by Gemini 2.5 Flash Lite

At its core, Phoenix runs on the Gemini 2.5 Flash Lite model — with the addition of custom-tuned capabilities.

Gemini 2.5 Flash Lite is a lightweight, fast, and efficient multimodal AI model from Google, designed for high-volume, high-frequency tasks. It can understand and process various types of data, including text, images, audio, and video, enabling it to tackle complex tasks that require understanding across different modalities. Currently Phoenix only supports text based inputs.

Data and Privacy

Privacy is Priority so here’s exactly how Phoenix handles your data:

Since Phoenix is powered by Google Gemini, you're also covered under Google's privacy policy for AI-related interactions.

We use basic tools like Google Analytics to understand general usage — things like:

Final Thoughts

Phoenix was created to feel real. Not in the uncanny, too-human way — but in the way that makes technology feel like a tool you actually want to use. It offers helpful, fast, and human-centered interactions — without the awkward AI detours or filler content.

From casual queries to more academic questions, Phoenix adapts to your tone and intent. It’s consistent, aware of its limits, and always focused on one thing: making sure you get the answer you need, in the way you need it.

And this is just the beginning.

Contributors: Aditi Chaturvedi, Akshay Pratap Singh