Vannus is an AI tool selection platform · elimination-first methodology · our scores aren't for sale — verified in code
What is Vannus Room?
The AI workspace that can't surprise you — not on the answer, not on the bill. A hard monthly cap means your bill never beats your subscription, and Pro+ routing is auditable with /why. Two models answer side by side; you pick the winner.
Predictable bill
The hard cap is a guarantee, not a limit — you're never billed past your subscription.
Auditable routing
On Pro+, /why shows which model answered and the exact rule that fired.
Two frontier models answer every prompt, side by side. You pick the winner; the loser is eliminated for that task.
Stop guessing which model to trust. Over 30 prompts you learn which model wins which job — that stack judgment compounds, and it’s the real product.
One subscription in place of ChatGPT + Claude + Perplexity. No tab-switching, no juggling logins.
Your bill never beats your subscription. Hard monthly cap — no metered surprises, ever.
Workspace persistence: every comparison searchable and tagged. Upgrade anytime — no data loss, no re-signup.
Under the hood
Multiple frontier models from leading providers, fired in parallel per prompt. Hard cap: $2/month of underlying LLM compute — you’re never billed past your subscription. The rotation evolves as the market does.
$14/mo$140/yr· $11.67/mo equiv.
Pro
The Personal engine, scaled up. Two models per prompt, you pick the winner — with 2.5× the monthly runway.
Same dual-model elimination as Personal — same engine, more room to run. Built for heavy daily use, long research threads, and code-review sessions that would hit the Personal ceiling mid-month.
The stack-judgment payoff compounds faster the more you run — you stop guessing which model to trust, sooner.
Your bill never beats your subscription. Hard monthly cap, 2.5× the Personal budget — still no metered surprises.
Workspace persistence; cancel anytime — access stays through your paid period.
Under the hood
Same multi-model engine as Personal, fired in parallel. The only difference is the compute budget: $5/month vs $2 on Personal — more runway, same no-surprise-bills guarantee.
$24/mo$240/yr· $20/mo equiv.
Most flexible
Pro+
Each prompt auto-routed to the best-fit model by published rules. Type /why to audit any decision.
The routing does the choosing for you. Instead of paying multiple vendors and tab-switching, every prompt goes to the model best suited to it — including specialized models you likely don’t subscribe to directly.
Nothing is a black box. The routing rules are published, and /why shows exactly which rule fired on any answer.
Larger context windows for long research threads and code review. Multi-model collaboration features included automatically as they ship.
Your bill never beats your subscription. Hard monthly cap, 3× the Pro budget — still no metered surprises.
Under the hood
Best-fit model per prompt from a rotation of frontier models, selected by rules you can read — and audit with /why. Hard cap: $15/month of underlying LLM compute. Pricing fixed at launch, reviewed annually. See the routing rules & current rotation.
$59/mo$590/yr· $49.17/mo equiv.
All paid plans are recurring subscriptions that auto-renew — monthly plans renew monthly, annual plans renew yearly — at the price shown, plus applicable tax, until you cancel. Cancel anytime from your account billing portal; cancellation stops future renewals and your access continues through the period you already paid for. See the Terms of Service (Section 12, Billing) for full details.
Pro fires two cost-efficient models side-by-side per prompt; you pick the winner, the loser is eliminated. Pro+ uses a single best-fit model selected by published routing rules — type /why after any Pro+ response to see exactly which rule fired. Routing rules and current model rotation are published at vannus.co/routing. Available models rotate as the market evolves — new providers are added, deprecated ones are removed, pricing-shifted ones are re-evaluated — without affecting your subscription terms.
Predictable bill, guaranteed in code
One flat price — your bill can never exceed your subscription. We absorb the compute cap so you never see a metered surprise.
Cancel anytime
Cancel anytime. Access stays through your paid period; no prorated refunds. Prompts never trained on. Full terms.
Payments processed by Stripe. EU/UK 14-day right of withdrawal applies (waived once paid compute is used). The free platform stays free.
Need a hands-on evaluation?
Vannus Concierge is a founder-led stack audit service. No junior handoffs. No templates.