Skip to content
Archwoven

ARCHITECTURE, CONNECTED.

Your system.
A clearer picture.

Turn descriptions and GitHub repositories into architecture diagrams you can explore, refine and share.

Free to explore. No account needed.
New AI requests are subject to shared usage limits.

FROM AN IDEA TO A SYSTEM01 / Architecture
THE DESCRIPTION

“A catalog app: customers call an API gateway; a catalog service serves items from PostgreSQL.”

Catalog architecture: Customer calls API Gateway, which calls Catalog Service, which queries Catalog DB in PostgreSQL. Each component is marked described.
Recorded example · arranged for this previewView SVG (opens in a new tab)
4 diagram typesCommit-pinned source evidenceSVG, PNG & JSON exportsLocal / private mode

A SHORTER PATH TO UNDERSTANDING

Start with what you have.

An idea in plain English. A public repository. A system you want to explain.

  1. 01

    Describe or import

    Describe your components and connections, or point to a public GitHub repository. Repository analysis reads source without running its code.

  2. 02

    Follow the connections

    Find a component, inspect a relationship or trace a path through the diagram. See what came from your description, source code or an inference.

  3. 03

    Make it yours

    Refine the diagram in words. Download an SVG, PNG or JSON file for a design discussion, project handoff or README.

ONE SYSTEM. DIFFERENT PERSPECTIVES.

The view that fits the question.

Architecture

See the components, services and data stores that make a system work.

Sequence

Follow a request through the participants, one interaction at a time.

ER / data model

Understand entities, fields and the relationships between them.

Component detail

Explore a component and the responsibilities inside it.

BUILT TO BE QUESTIONED

A diagram with reasons
behind its connections.

Architecture diagrams are useful when you can inspect their assumptions. Archwoven makes those assumptions visible.

Explore the code and measured results (opens in a new tab)
Evidence, where it exists
Source-supported claims cite files and lines at the analysed commit. Spring Boot and FastAPI have framework extractors; other repositories receive a bounded module and import map.
Structure before pixels
The model returns structured data. Schema and graph checks validate it before deterministic code renders the diagram. Validation does not prove that every architectural claim is correct.
Unknowns stay visible
Described, source-supported, inferred and proposed connections are distinguished. Coverage limits and unanswered questions travel with the diagram.
Choose where processing happens
Hosted generation uses an external AI provider. The local deployment uses Ollama, keeps data on your machine and enforces its outbound network boundary.

BEFORE YOU START

A few useful answers.

Can I generate a diagram from any GitHub repository?

You can submit public GitHub repositories within the import limits. Spring Boot and FastAPI get framework-aware extraction. Other repositories get a bounded map of source modules and imports, with gaps disclosed. Private repositories can be analysed from a local clone using the local deployment.

Does exploring this page use AI?

No. The example above is a recorded diagram. Opening the workspace also does not start generation. AI is requested when you submit a description, analyse a repository, refine a diagram or expand a component.

What happens when the free AI allowance is used up?

New AI calls pause with a message explaining the applicable limit. Viewing and downloading an available diagram do not make another AI call. Download files you want to keep; hosted recovery copies are temporary.

Can I use Archwoven privately?

Yes. Run the local deployment with Ollama. Local mode keeps generation on your machine. Setup instructions and limitations are in the repository README.

YOUR NEXT SYSTEM, MAPPED.

Make the complex clear.

Open workspace