One price. No procurement.
Linear, transparent, usage-based. The same plan whether you have ten users or ten million.
- 100,000 free API requests
- Unlimited flags and environments
- 5-year audit log retention
- Community support
- Hard usage cap — no surprise bills
- Everything in Hobby, and:
- Uncapped usage and unlimited teammates
- 5-year audit log with webhooks
- Percentage rollouts and scheduled releases
- Email support, response under 4 hours
Every capability, on both plans.
The full platform ships on the free plan. Pay-as-you-go lifts the usage cap, adds a few release controls, and upgrades support — that's the only difference.
A fraction of the invoice.
Equivalent feature set, at every volume. No annual contract, no committed spend.
Incumbent figures are illustrative, based on a typical ~$45 per million annual-contract rate.
Questions, answered.
What counts as a request?
One HTTP call to any /api/v1/* endpoint counts as a single request, no matter how many flags it evaluates. Batch-evaluate a hundred flags in one call and it still counts as one.
How does pricing work?
The first 100,000 requests are free, for the lifetime of your account and shared across every project. Beyond that it's $5 per million requests, billed in whole-million blocks. There are no tier jumps and no per-seat charges.
Do I need a credit card to start?
No. Create a project and ship your first flag on the free plan — no card, no sales call.
What happens when I reach the free limit?
Free accounts get a warning email at 80% of the quota and are hard-stopped once usage passes 125% — so you never get a surprise bill. Switch to pay-as-you-go to lift the cap and keep serving traffic uninterrupted.
Are there contracts, seats, or committed spend?
None. No annual commitment, no procurement cycle, and no per-seat pricing — invite as many teammates as you like on either plan. You only pay for the requests you actually make.
Do you offer volume or enterprise pricing?
Yes. For high-volume workloads, custom retention, or procurement requirements, reach out to contact@restflags.com and we'll tailor a plan.
Where does flag evaluation run, and how fast do changes propagate?
Evaluation is served by a stateless edge service handling roughly 25,000 requests per second, with a hot standby that takes over automatically if the primary instance fails — so evaluation keeps serving through failures, restarts and deploys. Flip a flag and evaluations pick up the change within about ten seconds, with a one-click global kill switch for emergencies.
Are there API rate limits?
Yes. The flag API allows up to 10,000 requests per second per API key and 50 per second per client IP, across every call — evaluations, events and metadata reads alike. Going over returns HTTP 429; cache results briefly on your side and you'll rarely come close.
Ship your first flag this afternoon.
100,000 API requests free. No card required.