KoltrixThe thread preview card in Koltrix: what each part tells you
Rest the pointer on a sender, summary or priority marker in Koltrix and a card shows the summary, importance, reply status and files. What each part means.
On this page(7 sections)
Opening a conversation to find out whether it matters is slow. You click, read, go back and repeat, and by row twelve you have lost your place.
Koltrix has a small preview card for exactly that moment. It answers the usual questions about a conversation, such as what it is about, whether it needs you and how long it has been waiting, without opening it. This post explains when the card appears, what each part says and how it behaves, so you can use it quickly and judge what it tells you.
When the card appears
The card opens when you rest the pointer on certain parts of a row in the inbox list for about two seconds:
- the sender (avatar or name);
- the AI summary line;
- the priority marker;
- the Needs reply chip;
- the attachment clip;
- the time.
Moving the pointer across the rest of the row does nothing, so scrolling down the list does not make cards pop up. Those parts show a help cursor so you can tell they are live. The card also works from the keyboard: focus a row and press p, and press Esc to close it. On touch screens it does not open on hover; you open the conversation instead.
It works in the inbox, in search results and in other lists such as Sent and Drafts. The section of the card that matches the part you pointed at is highlighted briefly.
What the card shows
From top to bottom:
Header. Who the conversation is with, when it last moved, which mailbox it came to or was sent from, and whether the latest message was sent or received.
Summary. One or two sentences about the conversation, then what the other person wants, and a deadline chip if one was mentioned. If the conversation has no summary yet, the card shows a quiet Summarize button instead. Hovering never asks the AI for anything on its own: the summary you see is one that already exists, or one you ask for with a click.
Status. Three quick facts:
- Importance: high, normal or low, with a short reason such as "known contact" or "deadline soon". It is a hint, not a ruling.
- Needs reply: whether the conversation is waiting on you, and for how long. Automated mail is not flagged. If you replied last, it says so.
- Category: which inbox tab the conversation belongs to.
Conversation. How many messages there are, how many came from you and how many from others, when it started and who is waiting on whom, with the people involved shown as small avatars.
Documents. If the conversation has attachments, they are listed with type, name and size, and clicking one opens the viewer. When the latest message is long, a short plain-text excerpt appears so you can see how it starts.
Labels and actions. The labels on the conversation, with a quick way to add or remove one, and a row of actions: reply, draft a reply, mark as done, archive, star, snooze and ask the AI about it.
Using the card to triage faster
A few habits make it pay off:
- Check "Needs reply" before opening. If it says the other side is waiting, handle it now. If the card shows you replied last, leave it. How AI classification works explains where the flag comes from.
- Use the waiting time. A conversation that has waited several days deserves attention before a fresh one, even if the new one looks more urgent.
- Clear what is done. "Mark as done" clears the Needs-reply flag for that one conversation. "Not needed" also stops flagging that sender in future, which is the right choice for automated notices.
- Look at the attachments. Often the question is whether a file came with the message.
- Ask for a summary only when it matters. The Summarize button uses an AI action, so save it for conversations you cannot judge at a glance.
A quick example
Say your inbox shows a conversation from a supplier with a small paperclip and a priority marker. Rest the pointer on the marker. The card opens with the Status section highlighted: importance is high because a deadline was mentioned, and Needs reply says the supplier has been waiting three days. The Documents section shows a PDF quote. This one needs you, so you open the quote, reply, and mark it done. The whole check took a few seconds, and you never left the list.
Now try a different row, a notification from a service you use. The card shows no reply flag and a low importance. You archive it from the card and move on.
How much to trust it
Summaries and importance labels are produced by AI, so they can be wrong. For a conversation with money, legal or deadline consequences, open it and read the original. When to trust AI thread summaries gives a checklist for deciding. If the tab or the flag looks wrong, correct it: Koltrix learns from that, and the feedback loop for misclassified email shows how corrections help.
Nothing on the card sends mail by itself. Draft reply creates a draft you review, and nothing leaves your mailbox unless you click send.
What it does not do
- It does not open the AI on every hover, so it costs nothing to browse.
- It does not change the conversation when it appears: it does not mark it read.
- It does not replace the reader. It is a fast first look.
Key takeaways
- Rest the pointer on the sender, summary, priority marker, Needs-reply chip, attachment clip or time for about two seconds, or press
pon a row. - The card shows the summary, ask and deadline, importance, reply status, conversation facts, attachments and quick actions.
- Hovering never calls the AI; Summarize is a click you choose.
- Treat summaries and importance as hints, and read the original when it matters.
Start with Koltrix
Your domain, one inbox, and an API that sends.
A team inbox where AI sorts and drafts (nothing is sent without your click), plus the transactional API and SMTP relay your product sends with. 7 days free, no card.
