Four flat-rate plans. Unlimited LLM compute on all of them — you only choose how fast it runs. Start in the dashboard, cancel anytime.
| Starter | Standard | Fast | Turbo | |
|---|---|---|---|---|
| Price | $9/mo | $39/mo | $99/mo | $399/mo |
| LLM compute | Unlimited | Unlimited | Unlimited | Unlimited |
| Model pool | Top-tier | Top-tier | Top-tier | Top-tier |
| Execution speed | Slower | Standard | Faster | Maximum |
| Scheduling | Shared pool | Shared pool | Priority | Highest priority |
| Batching | Heavy | Optimized | Reduced latency | Minimal latency |
| Trial | 3-day free trial | 3-day free trial | Money-back | Money-back |
| API keys | 1 | 1 | 1 | 1 |
Starter and Standard start with a 3-day free trial. At checkout we charge a $0.50 card verification — credited back on your first invoice — so your card is confirmed while you're present rather than failing silently on day 3. Cancel anytime during the trial and you pay nothing. One trial per person.
Fast and Turbo are billed at checkout because priority capacity is allocated to you immediately. Both come with a 3-day money-back guarantee instead of a trial.
Everything is managed from the dashboard billing page — no emails, no forms. Upgrades take effect immediately and are prorated, so you only pay the difference for the rest of the month. Downgrades apply from your next billing cycle. Cancelling stops future charges and your plan stays active until the period ends.
There is no per-token counter anywhere in the product. Agents retry, loop, and think — on per-token APIs every one of those is a small financial event. Here the price is the price: pick a speed tier, plug the API key into your agent, and stop watching the meter.
Yes — every plan is a flat monthly price with no per-token billing and no request caps. Instead of metering, we manage capacity with batching and dynamic throttling under load: higher tiers get faster execution and priority scheduling, but no tier ever hits a hard token wall. See the Fair Use page for exactly how this works.
Starter and Standard come with a 3-day free trial. You add a card at checkout, use the full plan for 3 days, and can cancel anytime before the trial ends without being billed. Trials are limited to one per person.
When you start a trial we charge $0.50 to verify your card while you're present — it prevents payment failures on day 3. The $0.50 is credited back on your first real invoice, so the trial effectively costs nothing.
No — Fast and Turbo are paid at checkout because they allocate priority capacity immediately. Both are covered by a 3-day money-back guarantee instead: if it's not what you expected, contact us within 3 days for a full refund.
Anytime, from the dashboard billing page. Upgrades are prorated immediately — you only pay the difference for the remainder of the month. Cancelling stops future billing and your access runs until the end of the paid period.
All plans use the same top-tier model pool — reasoning models from OpenAI, Anthropic, and xAI, selected per request by our routing algorithm. Plans differ in execution speed and scheduling priority, never in model quality. See the Models page for the current pool.
Speed. Every tier gets unlimited compute and the same models. Starter runs in a shared pool with heavy batching, Standard gets optimized batching, Fast adds priority scheduling and a higher-capacity pool, and Turbo gets the highest priority with minimal batching latency for sustained agent workloads.
More details on fair use and available models, or head to the dashboard to pick a plan.