Zuse

An AI coding agent that keeps your code on your machine.

A cloud planner sees file names and function signatures. A model on your own GPU writes the code, and your tests check every edit. Zuse lists every byte that leaves your machine.

What crosses the boundary of your machine Inside your machine: Zuse's own source code, 163 files, 7,945 lines, 316 kilobytes; the local index; the local model on your GPU; your tests. One line crosses the boundary to the cloud planner: 14 kilobytes of names and signatures, with no function bodies. The plan comes back. Your machine Your code 163 files 7,945 lines 316 KB Local index names, signatures, callers, tests Local model Gemma 4 on your GPU Your tests check every edit before it is kept Cloud planner via OpenRouter plan 14 KB names and signatures, 0 function bodies What crosses the boundary of your machine Inside your machine: Zuse's own source code, 163 files, 7,945 lines, 316 kilobytes; the local index; the local model on your GPU; your tests. One line crosses the boundary to the cloud planner: 14 kilobytes of names and signatures, with no function bodies. Your machine Your code 163 files 7,945 lines 316 KB Local index names, signatures, callers, tests Local model Gemma 4 on your GPU Your tests check every edit before it is kept 14 KB names and signatures, 0 function bodies Cloud planner via OpenRouter
Measured on Zuse's own source code: one planner request, as recorded in the ledger.

How it works

  1. Index on your machine

    Zuse builds a symbol index of your repository locally. It is never uploaded.

  2. Plan from an outline

    The planner gets your request and an outline: file names, symbol names and signatures. Function bodies stay behind.

  3. Edit on your machine

    A model on your GPU or CPU makes each change, one function at a time, in a format that has to parse before it is written.

  4. Check with your tests

    Every edit has to pass your type checker and tests. If it fails, Zuse retries with the real error, then rolls the edit back.

  5. Read the ledger

    Every request that left your machine is listed, byte for byte. Undo a whole run with one command.

Measured, including the misses

Numbers from our development machine. Results on the two reference laptops come before the preview opens.

What we measuredResult
Index a 100,000-line repository1.6 seconds
Re-index one changed file3 milliseconds
Planner request for a change to Zuse's own code14 KB, no function bodies
Local model finishes the task, given a correct plan, on an 8 GB GPU (Gemma 4 E4B)53 to 67% of 30 tasks
The same on a 4 GB GPU (Gemma 4 E2B)34% of 30 tasks

The small local model is good at single-function fixes and weak at changes that span several functions. We publish that so you can decide where it fits.

What to expect on your hardware

Your machineWhat Zuse does locally
GPU with 8 GB or more, or a Mac with 16 GB of memoryPlans and makes multi-step edits
GPU with 4 GB, or CPU onlyAnswers questions, plans, and makes single-function fixes

For safety-critical software teams

Zuse is founded by an engineer with two decades in automotive software. We are opening design-partner pilots for teams that work under ISO 26262 and MISRA.

  • Local mode sends nothing off your network, so code shared between OEMs and suppliers never moves.
  • Your MISRA checker runs on every edit, alongside your tests, and an edit that fails it is not kept.
  • Pilots include an evidence export for your safety case: for each change, the request, the plan, the diff, the check results and the ledger.
  • We support your tool-confidence assessment under ISO 26262-8. We do not claim certification; that judgement stays with your assessor.

Questions

What exactly leaves my machine?

In local mode, nothing. With cloud planning on: your request, an outline of names and signatures, and short summaries of results. Never function bodies, unless you approve a specific snippet. The ledger in the editor lists every request.

Which models does it use?

The local model runs through Google's LiteRT-LM on your GPU or CPU, Gemma 4 today. Cloud planning goes through OpenRouter, routed only to endpoints with zero data retention.

What does it cost?

Local mode is free and needs no account. During the preview, cloud planning uses your own OpenRouter key, and each request shows what it cost.

Which editors and languages?

VS Code first, including VSCodium and Cursor. TypeScript, JavaScript, Python and Go today; C and C++ next.