FAQ

Q: My Stacks view is blank, missing accounts, or showing resources in unexpected groupings. What should I do?

A: After updating your workspace subscriptions (adding/removing accounts or log sources), changes may take a few minutes to propagate to the Infrastructure View. If Stacks appears blank or incomplete on initial load, refresh the page and allow time for the architecture model to rebuild.

If specific accounts or resources are still missing after a few minutes:

Confirm your workspace subscriptions are correct. Verify that all expected AWS accounts and log sources are active under your integrations page.
Use tag-based filtering to isolate resources by team, product, or environment. If resources appear outside expected groupings, verify that tags are applied consistently across your infrastructure.
If the issue persists, reach out to [email protected] with your workspace name and the accounts/resources you expect to see.

For more information, see Using Stacks to Investigate Your Architecture.

Q: Are you using a Catio-trained model or a standard external LLM?

A: We use strong external LLMs, but that's not where the value is. The real value is across three key pillars, which compound over time:

Semantic Architecture Data: We build a continuously updated knowledge graph of your environment: systems, dependencies, decisions, ownership, and operational behavior over time. Anyone can read VPC flow logs or GitHub code. Nobody else has solved synthesizing that into semantically relevant architectural context. That's proprietary IP that gets richer the longer Catio is in your environment.
Reasoning Workflows, not one-shot LLM calls: When Catio answers an architecture question, it's not a single prompt. It runs a full reasoning workflow: planning steps, targeted tool calls, and multi-agent communications that replicate how a distributed team of architecture experts would actually solve the problem together. That's hard to replicate, and as we move toward proprietary architecture AI (LAMs and System Behavior Models), the gap grows further.
Architecture SaaS Workflows: Accumulated product depth, with Archie interfacing you to expert Stacks (for live architecture insights), Plans (for expert advice on optimizations and modernization), and Designs (for tailor-fit specs for software development, which are optimally coherent with your whole). These are essential for your team to materially improve your 1) time to architecture insight, 2) mean time to informed decision, and 3) daily control-to-execution loop, all of which are game-changers in how you reason on your system and keep it aligned over time.

A generic LLM gives general advice. Catio gives architectural intelligence grounded in how your system actually evolved, why it looks the way it does, and what constraints still matter.

Q: How do you keep client data secure? Are customers comfortable sharing infrastructure data with external LLMs?

A: We're SOC 2 Type II compliant, all processing runs strictly on AWS, and we have zero data retention / no training agreements with AWS Bedrock, which is our LLM provider. All our customers have put us through their enterprise security reviews and given us full access to their infrastructure after that.

For more detail, see our trust page: https://trust.catio.tech/