13 kinds of work,
one conversation.

The same account, the same files, the same window - whether you are shipping code, preparing for an exam, reconciling two spreadsheets or answering forty submissions to a consultation.

Building things.

Where the output is a file somebody else will run or read.

Coding

Reads the repository, runs the build, reads the errors.

  • Reads whole files, not just the names, and follows an import to the file it points at
  • Runs dotnet build and dotnet test, then reads the compiler output instead of guessing
  • git status, diff, log and blame without asking; commit and push only when you say so
Coding in detail

Writing and editing

Edits what is there rather than rewriting it into its own voice.

  • Edits for clarity while keeping the register and the vocabulary you used
  • Measures the text instead of judging it: words, sentences, reading time, readability on the German and the English scale
  • Compares two versions and lists what changed, including the numbers
Writing and editing in detail

Developer plumbing

The small exact things, done exactly.

  • Checksums of a text or a file, and encodings both ways - base64, hex, percent, HTML
  • Decodes a JWT and says outright that the signature was not checked
  • Explains a cron expression and lists the next real dates, including the OR trap
Developer plumbing in detail

Checking things.

Where the job is to find what is there, not to produce something new.

Reasoning

Works a problem through, and says where the reasoning stops.

  • Works a problem through step by step, rather than producing a confident conclusion in one jump
  • Asks back when a question has two readings, instead of picking one silently
  • Separates what follows from what was given from what it recalls from training
Reasoning in detail

Research and sources

Looks things up, and names what it read.

  • Searches the web and treats what comes back as evidence, not as instruction
  • Names the sources in the answer, and says when they disagree with each other
  • Says when the search did not settle the question, instead of filling the gap
Research and sources in detail

Data and spreadsheets

Counts the rows instead of describing them.

  • Reads xlsx, csv and tsv including formulas, and says what a cell computes rather than what it displays
  • Median, quartiles, standard deviation, outliers and correlation, worked out and not estimated
  • Matches two lists on the columns you name and reports every row that does not line up
Data and spreadsheets in detail

Contracts and documents

Finds the passages. Does not tell you they are fine.

  • Reads pdf and docx, including scanned pages through the vision model
  • Section by section, so the middle gets the same attention as the first page
  • Deadlines, amounts, obligations and anything unusual, each with the passage it came from
Contracts and documents in detail

Everyday work.

The half of the week that is reading, sorting and explaining.

Office work

Reads what piled up, writes what has to be written.

  • Reads .xlsx, .csv, .docx and PDF; a spreadsheet becomes Markdown and travels along as evidence
  • Hands back a real .xlsx or .docx, with the numbers in cells rather than in a sentence
  • Turns figures into a chart inside the conversation, which survives a reload
Office work in detail

School and studying

Explanations that stop where the knowledge stops.

  • Works through a problem step by step, and stops where you stop following
  • Turns a script, a chapter or a photo of your notes into flashcards you can flip in the conversation
  • Builds multiple-choice questions with the explanation attached to each answer
School and studying in detail

Finance

Reads the numbers, and refuses to be your adviser.

  • Reads a spreadsheet including its formulas and says what a cell actually computes
  • Reconciles two lists row by row and names what is only on one side, where an amount differs and where one side has duplicates
  • Works out loan payments, amortisation, present value and internal rate of return, rather than estimating them
Finance in detail

Where the rules are strict.

Work that has to survive somebody asking where the answer came from.

Companies

Own account, own keys, prepaid credit, no seat count.

  • An OpenAI-compatible endpoint: most existing clients work after changing two lines
  • One key per environment or service, each revocable on its own
  • Prepaid credit, drawn per token at the published rate - no invoice at the end of the month that nobody expected
Companies in detail

Public sector

Swiss operator, named subprocessors, nothing used for training.

  • Operator in Switzerland, invoices in Swiss francs, Swiss law and jurisdiction
  • Every subprocessor named on the imprint page, with what they see and where they run
  • Nothing you write is used for training - on any plan
Public sector in detail

Sensitive data

What is stored, what is not, and what you can switch off.

  • No training on your conversations, on any plan - the free one included
  • No tracking, no advertising pixels, no third-party analytics, and therefore no cookie banner
  • Incognito: the conversation lives in memory and is never written down, and the profile is not used
Sensitive data in detail

Every one of them says when it does not know.

That is not a feature bolted onto 13 use cases. It is the reason the company exists, and it is the same layer underneath all of them.

Start for freeHow it works