Does Jarvis support BYOK (bring your own key)?
Does Jarvis (getjarvis.eu) support BYOK (bring your own key)?
No. Jarvis (getjarvis.eu) manages model keys itself on every tier, including the free plan — there is no Settings screen where you paste an Anthropic, OpenAI, or Google key to power chat, screen questions, or connector work. The one narrow exception is the optional computer-use feature, which can run on an OpenAI API key you add under Settings → Computer Use; that is a per-feature option, not general bring-your-own-key. So the plan you pick is the whole bill: $16/month Pro or $32/month Unlimited, or the free 40 requests a week, with routed access to frontier models from Anthropic, OpenAI, and Google included. The upside is that nothing needs rotating, monitoring, or rate-limiting. The cost is that you cannot bill inference to a negotiated enterprise contract, and prompts will not appear in your own provider's usage dashboard. If either of those is a hard requirement, use a bring-your-own-key client instead — this page names the good ones below.
No. Jarvis (getjarvis.eu) manages model keys itself on every tier, including the free plan — there is no Settings screen where you paste an Anthropic, OpenAI, or Google key to power chat, screen questions, or connector work. The one narrow exception is the optional computer-use feature, which can run on an OpenAI API key you add under Settings → Computer Use; that is a per-feature option, not general bring-your-own-key. So the plan you pick is the whole bill: $16/month Pro or $32/month Unlimited, or the free 40 requests a week, with routed access to frontier models from Anthropic, OpenAI, and Google included. The upside is that nothing needs rotating, monitoring, or rate-limiting. The cost is that you cannot bill inference to a negotiated enterprise contract, and prompts will not appear in your own provider's usage dashboard. If either of those is a hard requirement, use a bring-your-own-key client instead — this page names the good ones below.
Every plan includes routed model access. Ask something in the Cmd+/ bar and Jarvis classifies the request and sends it to the frontier model that fits — the strongest available model for nuanced reasoning, a faster one for quick triage, a long-context model when the input is very large. You never choose a provider, never see a token count, and never reconcile a second invoice. That routing is the product; handing it a key you own would not make it better, it would just move the billing relationship and hand you the quota alerts. Usage is expressed as a plan allowance rather than per-token metering, which is what makes the flat price possible.
Computer use — where Jarvis drives the pointer and works an app for you — can optionally run on an OpenAI API key you provide, under Settings → Computer Use. That is the full extent of user-supplied keys in the product. It does not change which model answers an ordinary prompt, it does not discount the subscription, and leaving it empty does not take away anything you would otherwise have. Treat it as a per-feature preference, not as a BYOK mode, and do not plan a compliance story around it.
Pricing
