Question 1

Why did my prompt get routed to a different model than I expected?

Accepted Answer

The parser picks based on intent keywords + your selected mode + the active team. You can override by naming the model explicitly in your prompt (e.g. "Use Claude Opus to write…") or by switching modes / teams.

Question 2

What happens to my credits if a task fails?

Accepted Answer

Nothing — only successful tasks are billed. If three out of four tasks succeed, only those three are deducted.

Question 3

Can I use my own API keys?

Accepted Answer

Not yet — Terminal X uses pooled provider keys so you don't have to manage 15 different signups. BYO-keys is on the roadmap for the Business plan.

Question 4

Where's my data stored?

Accepted Answer

Sessions, commands, and results live in Supabase under your account. Memories are scoped to your user row. We never sell or share your data. Export everything anytime at Settings → Data & Privacy.

Question 5

Why are some brand tiles dim / not working?

Accepted Answer

A subset of the brands shown in the brain view are stubs — their parsing rules are wired but the provider integration isn't live yet. Those automatically fall back to a working alternative (e.g. Llama → Claude Sonnet, etc.) so you always get a real result.

Question 6

What's the inactivity timer?

Accepted Answer

If you're away from Terminal X for 30+ minutes and return, we auto-create a fresh session. Your previous one is still in the sidebar.

How Terminal X routes one prompt across 35+ AI models