The whole product

Everything Butler can do

Every tool a signed-in user can open, and the concrete jobs each one does. Grouped by the work rather than by the menu — the path under each name is where it lives in the app.

18 tools 4 groups EN · DE · FR

Access Butler

Writing & language

4 tools

The everyday surface. Everything here works on text you type, paste or attach.

/

Chat

A conversation with an assistant, with your own documents and the live web available to it.

Ask questions, get explanations, draft and rewrite text
Attach files — PDF, Word, Excel, images — and ask about them
Source the answer from one of your knowledge bases
Turn on web search for anything current, with sources listed under the answer
Edit Word, Excel and PowerPoint files in the chat — corrections, translations, new slides, formulas — and get the file back with its formatting intact
Ask for a new Word report, Excel sheet or PowerPoint deck with native charts, or turn a PDF into an editable Office file
Pick an effort level; the higher tiers reason before answering

/translator

Translator

Side-by-side translation of text or a whole PDF.

Translates as you type, with automatic language detection
Drop in a PDF and translate the whole document
Click any translated word for synonym alternatives
Swap the languages to translate back and check the meaning survived

/proofreader

Proofreader

A review pass over a document that returns suggestions you accept or dismiss one at a time.

Paste text or drop a .txt, .md or .docx file
Get suggestions grouped as grammar, clarity, engagement and delivery
Accept a fix to apply it, or dismiss it — saved automatically
Keep documents in a library and re-review after editing

/document-studio

Butler Studio

A collaborative document editor, with the assistant right inside the interface.

Create a Word-style document or a LaTeX project, or import a .docx, .md, .tex, .pdf or .zip
Select text to rephrase, fix the grammar, ask for suggestions or ask a question, then replace or insert
Discuss the whole document, or keep the discussion to the selection
Compile the LaTeX and read the PDF beside the source
Version history, sharing and export to PDF, Word, Markdown or a source archive

Your documents & what Butler makes of them

6 tools

Butler’s centre of gravity: build a document library, have it read, searched and cited for you — then put an assistant of your own on top of it.

/deep-research

Deep Research

A question in, a structured and fully sourced report out — Butler plans the subtopics, searches, reads the sources and writes.

Search the web, your internal knowledge bases, or both
Add the scientific indexes — arXiv, OpenAlex, Semantic Scholar, Crossref, Europe PMC — automatically or on demand
Two sliders: how many sources to read, and how long the report should be
Follow the run: planning, searching, gap review, writing
Numbered citations, a bibliography and a clickable link to every source used

/scholar-review

Scholar Review

An AI review of a PDF — paper, thesis or report — where every remark is anchored to the exact spot in the document.

Drop in a PDF and start the review
Read the remarks as cards — one click jumps to the highlighted quote in the PDF
Add a reviewer rubric, optionally
Pick the model and the language; switch the grammar and fluency checks on or off
Keep past reviews as sessions and come back to them

/modeltraining/knowledge

Knowledge bases

Your own document libraries. Butler indexes them so answers cite your files instead of guessing.

Create a base, then drop or drag in PDFs, Word files and text
Files are chunked and indexed automatically
Use a base in any chat, or run a deep research over it
Generate titles and descriptions automatically for what you upload
Share a base with colleagues, or attach it to an assistant

Knowledge → Explore

Database Explorer

Every knowledge base drawn as a graph, a map and an overview — so you choose what the AI should read before it answers.

See the people, organisations, places, topics and documents of a database and how they connect
Click a node — a client, a project, two reports — and ask: the answer reads exactly that, in full, with sources
Check every sentence of the answer against the passage it rests on
Ask for a Word report from the explorer chat: written section by section from the selection, ready to download
Start a deep research on a node, limited to its documents

See how a focused question beats a general one

You give the focus. The AI answers to the point.

A question asked of a whole database gets an answer assembled from whichever passages happen to match the words. Ask it about the right node instead — one composition, one client, the two reports that matter — and the AI reads those documents in full before it writes. You bring the judgement only a person has; the AI does the reading.

In the recording a research group asks for one figure. Asked of the whole database, the answer is a vague interim value; asked of the selected study, it is the exact measured result, cited to its page.

Human-enhanced

Your selection decides what is read. The answer is as good as your focus — and much better than a guess over everything.

Every claim traceable

Each sentence links to the passage it rests on; open it in the document with the words highlighted.

Word files from the chat

“Write a 10-page report on this client” becomes a real .docx — outline, sections and sources — ready to download.

Deep research on a node

Run a full deep research limited to the documents of the node you picked.

More about the Database Explorer

Knowledge → Add content → Website

Web Vacuum

Copies a whole website into one of your knowledge bases — pages, images and the documents it links to — and keeps that copy up to date.

Give a URL, a crawl depth and a page ceiling; stay on one domain or follow links outward
Pick which linked PDFs, Word and Excel files to fetch too — or download them all as a ZIP
Browse the offline copy with its menus and accordions still working, even without a connection
Ask questions to the entire website in any chat, with the page each answer comes from
Re-vacuum it automatically every week or month: only what changed is downloaded again, and every version is kept with its differences

See it vacuum the TU Wien website

A website you own a copy of — current, searchable, offline

Point the Web Vacuum at a site — here the TU Wien institute pages, two levels deep. It reads every page like a browser would, keeps the images and the layout, and lists the documents the pages link to so you choose which ones come along.

The result is yours: a copy on your own server that answers questions about the whole site at once, and that you can still open when the site changes, goes down or you are offline.

Ask the whole site

Attach the base to a chat and ask — the answer cites the exact page, not a search snippet.

Stays up to date

Set a weekly or monthly re-vacuum. Unchanged pages are skipped, changed ones re-indexed, removed ones taken out of the answers.

Every version kept

Each re-vacuum is a version; open a changed page to see before and after side by side.

Works offline

Menus, tabs and FAQ answers still open in the copy — without running any of the site’s scripts.

/modeltraining

Assistant Training

Build your own chat assistant: give it a role and connect it to your knowledge bases.

Build a custom assistant, give it a role, and attach knowledge bases to it
It answers from your documents only
Share it with your group or keep it private
It shows up in the chat model picker like any other assistant

Meetings & records

3 tools

Recording, speaker-labelled transcription and an editable report — from a file, your microphone, or the sound of an online call on this computer.

/Meeting_transcription

Meeting transcription

Turns speech into a written transcript, then into a report you can edit and export.

Drop in an audio or video file — MP3, WAV, MP4, M4A, FLAC, OGG, WEBM — or paste a link to a video or a podcast and let the server fetch the sound
Record the room from your microphone, with pause and resume
Record an online call — Zoom, Teams, Meet, Webex — by capturing your microphone and the computer’s sound on two channels: Butler never joins the meeting
Diarisation: who said what, with a speaker count you fix or let it detect
Optional live transcription, which also gives you live translation
Edit the transcript, produce a summary or a report, and export to PDF or DOCX
File meetings in folders and share them with your group

See a meeting become a transcript and a report

Every word, every speaker, a finished report.

Drop in a recording, record the room, or capture an online call. Butler writes down who said what, names the people it knows, and turns the meeting into a report you can edit and export.

In the recording, a two-minute project meeting is transcribed with its three speakers — on Butler’s own GPUs, without the audio leaving the server.

Who said what

Speakers detected, then named — once a voice is in a dictionary, automatically.

Correct in place

Click a sentence to fix it, a name to rename that speaker everywhere.

A real report

Summary, decisions, actions and open questions, in your own template.

Word or PDF

Export the report alone or with the full transcript.

More about meeting transcription

/Meeting_transcription/templates

Meeting Report Templates

Manage a set of templates that control what goes into a meeting report and how it reads.

Define what belongs in the report, in substance and in form
Fix the AI model and the report language, or leave both on automatic
Pick a template when you transcribe; the report is written as soon as the transcript is done
Manage your meeting templates and share them with your contributors

/Meeting_transcription/dictionaries

Speaker Dictionaries

Teach Butler the voices of your group so transcripts carry real names instead of “Speaker 1”.

Identify a person from several voice samples
Butler matches the speakers against the dictionary on every new transcript
Automate transcripts that name the people in the room

Making things

5 tools

Output you can put in a talk, a paper, a poster or the workshop — slides, animations, drawings, PDFs and 3-D parts.

/presentation

Presentation Creator

Turns a paper, a thesis or a report into a compiled LaTeX Beamer deck — figures placed, slides written, PDF built.

Drop in a PDF, Word, LaTeX, Markdown or text file, or simply write an instruction
Set the talk length, the audience, the language, the theme and the density
Butler describes your figures, plans the talk, writes every slide and places the figures from the paper
Export the PDF or the LaTeX source, which opens in Overleaf or Butler Studio

/animation

Animation Studio

Make your own animations: Butler writes a scene with Manim — the library behind the 3Blue1Brown videos — renders it and hands you the video.

Put animations into your presentations
Choose the length, the quality (draft or 1080p) and the format (MP4 or GIF)
Iterate with Butler until the animation is the one you want

/pdf-handling

PDF Handling

Six utilities for working with PDFs.

Edit — change the text of a PDF invisibly
Reorder and merge — move pages around or join several PDFs
Compress — shrink a file so it fits in an email
Protect and unlock — add or remove a password
PDF to images — every page as PNG or JPEG
OCR — pull the text out of a scanned document

/excalidraw/boards

Excalidraw Boards

A flipchart for sketches, diagrams and visual notes, saved as boards.

Draw with shapes, arrows, freehand strokes and text
Boards save themselves and show up as thumbnails in your list
Keep a board private or share it with your group

/cad-studio

CAD Studio

Describe a part in plain words and get a parametric CAD model back — exact STEP for the workshop, STL for the 3D printer.

Describe the part with its dimensions in millimetres, and optionally attach a sketch, a drawing or a photo
Butler writes a parametric build123d (Python) script; the OpenCascade kernel builds it in an isolated sandbox with no network
If the kernel reports an error, the real message goes back to the model, which repairs the script — up to three times
Download STEP (SolidWorks, Fusion, FreeCAD, Onshape, CNC), STL (3D printing) or GLB (web)
Exact measurements on the B-rep geometry: size, volume, surface, mass for common materials, centre of mass, valid-solid check
Refine in plain words (“make the walls 3 mm”); every run is a version you can restore
A 3-D viewer with iso, top, front and right views, feature edges and a section plane

See a brief become a STEP file

Describe the part. Get the STEP file.

Type what you need — a sample holder, a fixture, a test cell — and attach a sketch if you have one. Butler writes a parametric CAD model, builds it with the OpenCascade kernel and hands you STEP, STL and the exact measurements of the solid.

In the recording, a part is described in a few lines, built, cut open with the section plane and measured — without a CAD program open.

Real CAD files

STEP for CAD and CNC, STL for 3D printing, GLB for the web.

Exact measurements

Size, volume, mass and centre of mass, measured on the solid, not the mesh.

Repairs itself

The kernel’s real error goes back to the model, which fixes the script.

Every run a version

Refine in plain words, restore any earlier version.

Things you do inside a chat

These are not entries in the dock — they are behaviours available in any conversation.

Attach a document

Drop a PDF, Word, Excel or image into the message box and ask about it. Several files at once is fine.

Cite your own library

Type # and pick a database. The answer quotes your files and shows which passage it came from.

Search the web

Toggle it on to read live pages, or leave it off — Butler turns it on by itself when a question clearly needs current information.

Word, Excel and PowerPoint

Ask for a change to an attached document, workbook or deck and get the file back with its formatting intact — edited structurally, not regenerated. Or ask for a new one.

Watch it think

Higher effort tiers reason before answering, and the reasoning trace is there to open if you want it.

See what answered

The generation-info hover on any message names the model that actually served it and how long it took.

All of it, on your own hardware

Every tool on this page runs on machines you control. Try the free ones anonymously, or open Butler if you already have an account.

Access Butler

Five of these run anonymously, with no account, on our free lane.