Pricing

The easiest yes in your stack.

One price. No seats, no platform fee, no token markup. We only make money on the spend that flows through us — and we prove you saved far more than we charged.

Standard
2%of managed spend

Everything included. Cancel anytime. Start free — you’re only billed once real spend flows through.

Start free →
  • Unlimited seats & workspaces
  • Every provider, one endpoint
  • Budgets (team / app / agent)
  • Policy engine + auto-stop
  • Full observability & audit
  • Cost-optimization + recommendations
  • Self-serve access management
  • SSO (OIDC) + RBAC
  • Agent kill-switch
  • Self-hostable

What would it cost you?

$50,000/mo
Byoky fee (2%)$1,000/mo
Typical spend cut (20–40%)$10,000$20,000/mo
Net savings after our fee$9,000$19,000/mo
Savings from cheaper-model routing, response caching, and budget caps — measured against a real no-Byoky baseline.
Why this model

Aligned with you, not against you.

❌ Per-seat

Gets more expensive every time you hire — punishes exactly the growth you want. A cost-control tool shouldn’t be a growing cost.

❌ Token markup

The reseller earns more when your bill goes up. That fights the entire point of buying a tool to cut spend.

✅ 2% of managed spend

We win only when spend flows through us — and we’re simultaneously shrinking it. Prove the cut, justify the fee.

FAQ

Straight answers.

Is there a seat fee or a flat platform fee?

No. 2% of managed spend is the only charge. Adding people or workspaces costs nothing — a cost-cutting tool shouldn’t get more expensive as you grow.

Do you mark up tokens?

Never. You keep paying your providers directly on your own accounts (and keep your committed-use discounts). We meter what flows through and bill 2% of it.

What counts as "managed spend"?

The actual provider cost of the requests routed through Byoky in a billing period — not budget you allocate, and not spend that fails open during an outage.

What if Byoky saves me less than 2%?

The console measures savings against a real no-Byoky baseline. If the optimizer, caching, and caps don’t clear the fee, you’ll see it — and you can turn the fee-bearing features off. In practice a 20–40% cut dwarfs a 2% fee.

Can we self-host?

Yes — the full stack ships as Docker/Helm for on-prem or your own cloud, with your own KMS and IdP.

Start governing your AI spend.

Connect a provider key, point your SDK at Byoky, watch the spend — in under 10 minutes.