Skip to content

Frequently asked questions

Everything you need to know about arch0.

No. ChatGPT produces free-form text that varies with every prompt. arch0 uses a schema-enforced pipeline that generates a fixed set of 11 validated deliverables. The outputs are deterministic — same inputs, same structure. You get a structured blueprint, not a chat transcript.

Yes. The export zip contains standard markdown files, Mermaid diagram source files (.mmd), and JSON configuration. These formats integrate with any development workflow — paste the Mermaid code into Cursor, import the markdown into your wiki, or feed the manifest.json into your CI pipeline.

Iterate within the same project. When you update requirements, the agent regenerates only the affected deliverables and tracks the differences. You can see exactly what changed between versions using the built-in diff view.

Cost estimates are based on published cloud provider pricing (AWS, Azure, GCP) and calibrated against common architecture patterns. They provide a realistic monthly estimate for the services recommended in your architecture. Actual costs may vary based on usage patterns, reserved instances, and regional pricing. Estimates should be treated as a planning baseline, not a quote.

No. arch0 generates architecture recommendations and documentation — it does not execute code, provision infrastructure, or deploy services. The deployment guide includes infrastructure-as-code recommendations, but execution is up to your team.

Yes. You can delete individual projects at any time from your dashboard, which removes all associated messages and deliverables. For full account deletion, contact us through the contact page. Account deletion is processed within 30 days and removes all data permanently. You can also export all your data before deletion.

Your data is used only to provide the service — generating architecture recommendations and deliverables within your projects. We do not use your data to train AI models. See our security page for details on data handling, encryption, and retention.

arch0 uses Anthropic's Claude model family. The agent runs a multi-turn loop with tool access, meaning it can search knowledge bases, generate diagrams, and produce structured deliverables — not just generate text.

Yes. arch0 supports four project types: greenfield (new systems), migration (moving between architectures), review (evaluating existing architecture), and general (open-ended architecture questions). The agent adapts its questions and deliverables based on the project type.

API access is available on the Team plan. It provides programmatic access to project creation, message sending, and deliverable export. Documentation is available after upgrading to the Team tier.

Still have questions?

Reach out and we will get back to you.