Dotrai: A Free AI Chat That Shows Its Working

Dotrai: A Free AI Chat That Shows Its Working

Dotrai is our free AI chat. Most AI chats give you a confident paragraph and leave you to decide how far to trust it. Dotrai shows how it got there. It searches the live web and links the sources it used, it runs the code it writes instead of guessing what that code would print, and it lets you watch its reasoning as it happens. You get the answer and the means to check it.

Why the working matters

A wrong answer from an AI chat rarely looks wrong. It is fluent, well formatted and confident, exactly like a right one. Confidence tells you nothing, so the only real way to decide whether to rely on an answer is to see how it was reached.

That is the idea Dotrai is built on. The answer is only half of what it gives you. The other half is the working behind it, shown in a form you can follow. If the reasoning skips a step, you can see the gap. If a claim comes from a source, you can open the source. If code was involved, you can see what it actually printed.

What "showing its working" means

  • Visible reasoning. The steps behind an answer appear while Dotrai is working, not as a tidy summary afterwards. A summary of reasoning is just another paragraph asking to be trusted.
  • Live web search with sources. When a question depends on current information, Dotrai looks it up, and the answer carries links to the pages it used so you can follow a claim back to where it came from.
  • A code sandbox. When Dotrai writes code, it runs it in a sandbox. The output you see is the real output.
  • Files and images. Upload a document or an image and ask about what is in it.

Current, not remembered

A language model knows what it was trained on, and that knowledge stops at a date. For a lot of everyday questions this is not a small limitation. What something costs now, what changed last week, what the latest version of a tool does: a model answering from memory gives you something that was true once, in the same confident tone as everything else it says.

Dotrai searches for those answers instead and tells you where it found them, so you can see for yourself how recent they are.

Code that has actually run

The sandbox is the clearest test of the whole approach. A model can write a plausible program and an equally plausible description of what that program prints, with nothing connecting the two. They differ more often than you would expect, and they differ silently. Running the code settles it: either the program printed that result or it did not.

This is useful well beyond programming. Calculations, unit and data conversions, and quick checks on a table of figures are all more reliable when the code doing them has really run instead of being imagined.

History and incognito chats

Conversations are saved to your history so you can come back to them. When you would rather not keep something, start an incognito chat, and nothing from that conversation is written to your history.

We made incognito a visible choice rather than a setting hidden in a menu, because many of the questions worth asking are ones people would rather not keep on a list: health, money, legal questions and personal matters among them.

Who it is for

  • People who want to check an AI answer rather than take it on trust.
  • Students and researchers who need to see where a claim comes from.
  • Developers who want code that has actually been run.
  • Anyone whose question depends on information newer than a model's training.

How to start

  1. Open dotrai.com and ask a question. You do not need a card to start.
  2. Watch the reasoning, the searches and any code it runs as they happen.
  3. Open the sources linked in the answer and check the claims that matter to you.
  4. Keep the conversation in your history, or use an incognito chat and leave nothing behind.

Where it is now

We started building Dotrai in March 2025 and spent most of the time since developing it rather than promoting it. It is open to everyone now, free and without a card, and we are still working on it. Like everything we build, the thing it exists to do is not the thing you pay to find out whether it works.

Frequently asked questions

Is Dotrai free?

Yes, and you can start without a credit card.

What does "shows its working" mean?

The reasoning behind an answer is visible as it happens, the searches it ran are listed with their sources, and code is executed rather than predicted. You can check the answer instead of taking it on trust.

Does it really run code?

Yes, in a sandbox. The output you see is the result of running the code, not the model's guess at what it would print.

Can it answer questions about recent events?

Yes. For questions that depend on current information it searches the live web and links the pages it used, so you can see where the answer came from.

What is an incognito chat?

A conversation that is not saved to your history. It is useful when you are asking about something you would rather not keep.

Try it at dotrai.com.

Previous Post Next Post
//