One question across every source in your projects.
The AI assistant searches by meaning rather than by keyword, across tasks, comments, documents, documentation pages, notes, conversations with their analyses, and the source code of your projects. It runs as a narrow track beside the page you are on, and it belongs to no single project. Because it all sits in the same body of data, one question can hit several sources at once.
- 01 Answers with their sources cited, linking to the ticket, the page or the conversation
- 02 The source code is indexed as structure, not as a text file, which makes questions about the implementation answerable
- 03 Project permissions apply in the chat too. No answer comes out of an area someone may not read
- 04 The AI assistant stays open while you keep clicking. The history is yours alone
- Are the passwords in our software encrypted?
- Does our signup form have separate fields for postcode and city?
- What did we promise the client about the interface, and what does the code say about it?
How this works
You ask the question you actually have, in full sentences. The search is by meaning, not by keyword, across tasks, comments, documents, documentation pages, notes and released meetings.
Under every answer sit the sources it came from, clickable through to the ticket, the page or the meeting. An answer without evidence is a claim, and claims help nobody on a project.
The source code is read as structure, not as a text file. Questions about what is actually implemented can be answered against the repository rather than against what someone wrote about it.
The AI assistant stays open beside you while you move from the task to the meeting to the note. The next question knows the previous one, so you do not have to explain the context again.
What it does not do
Two limits the chat has on purpose.
- It reads, it does not write. The chat creates no tasks, changes no status and comments on nothing. What happens on the project happens because a person did it.
- It knows only what has been released. A meeting you have not released appears in no answer, not even in fragments.
What people ask at this point
- How do I know the answer is right?
- Because it names its sources and you can click them. Checking then takes seconds instead of half an hour, and an answer without sources tells you at once that you need to look yourself.
- Do I have to switch the AI on for this?
- Yes. As delivered it is off, and without your release no content leaves the system. Which provider does the computing you decide in the same step.
- What about source code I do not want to share?
- The chat reads what the project reads from the connected repository. Connect none, and there is no code in the chat.
Look at it
rather than take our word for it.
Productive in under 5 minutes. No sales loop.
Particular data protection or compliance requirements? Talk to us. For cases like that we usually find a way.