What counts as a "board"?
A board is defined by its outline, BOM, and pin count at the time of upload. You can iterate on that board as many times as you want, including changing floor plans, stack-ups, and component positions, as long as the BOM stays within 10% of the original. If the BOM changes by more than 10%, that's a new board.
What if I pre-route part of the board?
You only pay for unrouted pins. If you route the RF section, the high-voltage traces, or the BGA fan-outs before uploading, those pins aren't counted. Quilter routes around your existing work and charges only for what it needs to compile.
How many times can I iterate on one board?
As many as you want. Upload, download, adjust in your ECAD tool, resubmit. Try different stack-ups, different floor plans, different component placements. There's no limit on iterations and no extra cost. We want you to get a board you're ready to fabricate.
Can I run multiple jobs at the same time?
Yes. You can run parallel jobs within the same board: different stack-ups, different room configurations, different trace widths. Submit them before you leave for the day and review candidates in the morning.
Who owns the output?
You do. The board files are yours. Download them, fabricate them, modify them. There's no proprietary format and no lock-in. The output is a standard board file that goes right back into your ECAD tool.
What does guided onboarding include?
On your first board, a Quilter engineer schedules an hour and a half to walk through the submission process with you: uploading your files, reviewing the circuit comprehension report, setting up the floor plan, and clicking submit. After that, we check in regularly as you iterate to make sure you're getting the outputs you want.
What's the difference between deployment options?
The AI is identical. Cloud is fastest to start with the lowest cost. Private cloud runs in your own AWS account for tighter data control. Both use per-board pricing; the self-hosted option has a higher cost and might include additional fees.
What does the self-hosted deployment look like?
Quilter is a Kubernetes-native application deployed into your existing cluster via a Helm chart. If your organization runs a Kubernetes environment such as Amazon EKS, Quilter fits directly into that infrastructure.
We provide the Helm chart and container images, and your team deploys and manages it using the same tools and workflows you already use for other cluster workloads.
At steady state, Quilter’s baseline services consume on the order of a few dozen GB of RAM. When design jobs are submitted, Quilter creates additional Kubernetes jobs to handle the workload. Your clusterinfrastructure is responsible for scheduling and scaling to meet that demand, giving you full control over autoscaling behavior and resource allocation. Design jobs include a GPU workload, so your cluster will needaccess to NVIDIA GPU resources.
We recommend an outbound network connection where possible, as it simplifies updates and enables more responsive support. That said, Quilter can operate in fully air-gapped environments if your security requirements call for it. For authentication, Quilter integrates with your existing identity provider, so user management stays within your organization’s existing access controls.
