Pricing model
BYOK degressive markup. Your providers bill you directly. 9-12.5% on Scale.
HiWay pricing is transparent: your providers bill you directly for inference at their published rates, and HiWay applies a markup % on your monthly API spend. No hidden per-token cut, no flat subscription - one transparent percentage that decreases automatically with volume.
Free plan to start - no credit card
Every new account gets immediate access to the Free plan, no credit card required. Scale kicks in automatically as your API spend grows.
How BYOK works
- Create an account - immediate Free plan access.
- Plug your own provider API keys in Settings → Providers.
- HiWay routes requests using your key - the provider bills your card at their published rates.
- HiWay measures your monthly API spend and applies the markup for your volume tier.
Plans
HiWay2LLM Plans
| Plan | HiWay Markup | Monthly API Volume | Features |
|---|---|---|---|
| Free | - | Basic features | Routing, stats, 1 API key |
| Scale | 12.5% | < $500/month | All features |
| Scale | 11% | $500 - $5,000/month | All features |
| Scale | 10% | $5,000 - $20,000/month | All features |
| Enterprise | 9% | $20,000 - $50,000/month | Dedicated SLA, annual contract |
| Custom | Negotiated | > $50,000/month | Custom terms, dedicated support |
Automatic degressive markup on Scale
The markup is adjusted each month based on your actual API volume - no action required. It applies to your API spend (pre-routing). Auto-reload automatically tops up your HiWay wallet when the balance drops below a configurable threshold.
How is the markup calculated?
The markup applies to your monthly API spend - what your providers billed you this month. You also benefit from routing savings (auto-downgrade to cheaper models) which reduce your provider bill and therefore your markup base.
Included in every plan
- Smart routing across your enabled providers
- Guardian - anti-loop rules per workspace
- Provider fallback on upstream errors (max 2 retries)
- Budget Control to cap your upstream BYOK spend
- Qdrant semantic cache
- A/B Experiments - benchmark models on real traffic
- Opt-in PII masking
- CORTEX - on-board AI (triage, workspace insights, strategic advisor)
- Open-source CLI, Python and TypeScript SDKs
- Dashboard, usage logs, savings vs Opus 4.7 baseline
Enterprise: dedicated SLA + negotiated markup
Above $20K/month API spend: negotiated fees, annual contract, dedicated SLA. Contact [email protected].
Wallet at zero? Service keeps running
If your HiWay wallet hits 0, requests automatically switch to passthrough mode, BYOK direct, markup = 0%, smart routing suspended. You get a 72-hour OR 100,000-token grace period to top up, with a warning email at 50%. Beyond the cap, new requests return HTTP 402 until your next top-up. See the Passthrough mode concept page for the full behavior.