What Literess can do
Literess is your in-app translation assistant, and she does the work rather than only describing it. Tell her what you’re working on and she’ll set it up: build a glossary, assemble a workflow, move a document along your pipeline. Ask her where something is and she’ll take you there.
On this page
Where do I find Literess?
Inside a document she’s a left-docked sidebar in translation mode — document-aware, able to comment on blocks, edit glossaries, and run tools. Outside a document she’s a bottom-right widget in support mode — product and account help. She switches mode mid-conversation when the topic changes.
Can Literess do things for me, or only explain them?
Actions that change something are gated by a confirmation chip you approve first — and anything that spends words shows an estimate. She can build, edit, and run workflows, move a document to a CRM status, create a new document, invite a teammate by email, and set workspace defaults.
Workflows are a special case worth calling out: if you have one open in the builder (including a template you just picked), ask her to change it — “make the find step look for swear words,” “add a proofread before the export” — and her edit drops straight onto your canvas with an Apply button. No “which one?”, and you save it yourself with the builder. Building from scratch or editing a saved workflow without the builder open uses the usual confirm-first chip instead.
She can also drive the UI instantly (no chip): open a panel (glossary, styleguide, comments), take you to a page (CRM, billing, settings), or open a specific document. So “set up a translate-then-export workflow,” “invite [email protected],” and “open the glossary for me” are things she does — not just describes.
What if Literess proposes several changes at once?
Several proposals from the same turn arrive as one approval card instead of a stack of separate chips. The card lists what she wants to do, and Approve all runs every action on it in one click.
Each row keeps its own reject button, so you can approve the set and drop the one item you disagree with. Cancel all turns everything down. Nothing runs until you approve, exactly as with a single chip, and she confirms the whole batch in one reply.
Can Literess help me while a dialog is open?
Yes, on a desktop screen. Open her and she stays beside a dialog instead of vanishing behind it, and both stay usable. The page behind dims, clicking that dimmed area no longer closes the dialog (use its own Close or Cancel), and Escape closes whichever of the two you are typing in. If she was closed when the dialog appeared, her header pill in the editor and her bubble everywhere else stay clickable, so reaching for her does not cost you the screen you were on. On a phone the dialog still takes the whole screen, so there is no side by side there.
She can also read what the open screen holds and change its settings for you. On a merge screen that is which glossary you picked, which side wins a particular clash, and the counts in the preview. On a build or generation screen it is the name a run will create, the target languages it covers, and which suggested terms or rules are kept. Ask her to change any of it and she changes it, in words instead of clicks.
She never presses the button that commits. Starting a merge, starting a generation, applying a review: those stay a click you make. Nothing she does to a form spends words.
What can I attach to a Literess chat?
Paste, drop, or attach up to five files per message. She sees images natively (“what does this screenshot say,” “translate this menu”) and reads the text of documents (PDF, DOCX, TXT, MD, HTML) you send for context.
Can Literess look things up on the web?
Yes. When a question calls for it, she searches the live web as part of answering: a current fact, how a phrase is really used in the target language, a term of art, a cultural reference. There is no button to press and nothing to set up; a “Searching the web” pill shows while she looks, with the query she is running.
The pages she drew on appear as numbered sources under her reply, so you can check her reading yourself. Searching never spends words.
Can I just ask Literess to translate my document?
Yes, though the way she does it is by building and running a translate workflow for you — there is no one-shot translate button she can press. Like any run that spends words, it is gated by a confirmation chip showing the estimate before anything happens.
Can Literess find and read my documents?
Yes. Literess can search your documents and past chats by meaning, then read a document’s text or answer a focused question about it. It is how she resolves a vague reference — “my piece about the festival”, “where we discussed that term” — before she acts.
Why does Literess message me first, and how do I stop it?
Those are check-ins: when something fails (an import breaks, a run dies, your words run out) or you seem stuck, Literess can send one short message on her own. They are rare, capped per day, and never interrupt an active chat. Turn them off at Settings → Notifications → “Literess check-ins”, or just tell her to stop messaging you first — she flips the same setting.
Does Literess reply in my language?
Yes — Literess writes in whichever language you have set for the Transept interface (the Languages button in the sidebar, or Settings → Account). That is separate from a document’s translation: changing your interface language doesn’t change which languages a document is translated between.
How big a file can I send Literess?
Up to 10 MB per attachment, and up to five attachments per message. If a document’s text runs past roughly 80,000 characters she sees a truncated version — she will tell you when that happens, so you can send a specific section instead.
Does approving several actions at once cost more than approving them one by one?
No. Words are charged for the work itself, so a run costs the same whether you approved it alone or alongside others on the same card. Actions that only change a setting (a CRM status, a workspace default) never spend words at all.
Can Literess start the run once she has filled in a dialog for me?
No. She sets the dialog up: picks the glossary to bring in, answers a conflict row, chooses where a generated result should be saved, switches a suggested term off. The button that starts the work, or applies a review, is yours to press. Adjusting a form costs nothing, so there is no bill until you press it.
Does asking Literess something close the dialog I have open?
No. On desktop she sits beside it, so you can ask a question or change a setting without losing the screen you were on. Clicking into her no longer counts as clicking outside the dialog, and Escape closes whichever of the two has your cursor in it.
Does Literess answer questions about Transept itself from the web?
No. For facts about the product (plans, prices, what a feature does), she answers from her built-in product knowledge, and web results never override it. That keeps her from repeating a stale third-party page about Transept back to you as if it were current.
Still stuck? Ask Literess in the app, or write to [email protected].