Simple, transparent pricing

Frequently asked questions

Is there a free plan?

Yes. Glider (Free) is for side projects, hobby apps and trying GraphPilot out, with no credit card required. It is not meant for production traffic: it comes with no support commitment, and once its monthly request unit allowance is used up, edge caching and request logging pause until its monthly window resets. The paid plans Jet (Starter) and Airliner (Business) add larger allowances, more services, longer retention and deeper Insights. Team seats are unlimited on every plan.

When does billing start?

Glider (Free) costs nothing and needs no credit card. A paid plan is billed from the moment you order it in the portal: its price is charged at once, in advance, monthly or yearly. Usage above the plan's allowance is billed monthly in arrears, also on a plan billed yearly.

Can I change plans later?

Yes. You can start on Glider (Free) and move up to Jet (Starter) or Airliner (Business) as your GraphQL traffic and team needs grow. Jumbo (Enterprise) adds custom pricing and volumes, and a custom SLA and support.

How many organizations and custom domains does a plan include?

Both are counted per billing account, across every organization it funds. Glider (Free) funds 1 organization and includes no custom domain, Jet (Starter) 2 organizations and 2 custom domains, Airliner (Business) 3 organizations and 5 custom domains. Jumbo (Enterprise) includes at least 5 custom domains, and its contract sets the organizations and any further domains. An account that already funds more organizations or holds more custom domains than its plan includes, for example after moving to a smaller plan, keeps them; only a new one is refused.

Is there an audit log?

Yes, on every plan. Each organization has an audit log in the portal that shows who changed what and when: services, deployments, purges, history resets, signing keys, custom domains, plan and billing changes, members and API keys, settings and the acceptance of the terms. It names the member, the API key, GraphPilot support or GraphPilot itself that made the change; member and API key changes come from WorkOS, which manages them, and show WorkOS as the actor. It records no request contents and no secrets. Every member can view it, whatever their role; API keys never can, and no one can change or delete an entry. It reaches back 1 day on Glider (Free), 30 days on Jet (Starter), 180 days on Airliner (Business) and 365 days on Jumbo (Enterprise), and it is stored in the EU.

What happens when Glider's request units run out?

Your API keeps working: requests are still authenticated, signed and forwarded to your origin. Edge caching and request logging pause together until Glider's monthly window resets or you move to a paid plan. The window starts when the account moves to Glider, at sign-up or after a paid plan ends, and repeats monthly from there: requests served in that time are counted but not recorded, so they do not appear in Insights and an upgrade does not bring them back. Glider has no separate Insights allowance: telemetry is recorded only while the request units last; once they are used up, the account serves in lean mode and records no telemetry. Upgrade before the allowance runs out if you need a record of every request. Glider is offered under fair use; see below.

Does Glider ever stop serving?

Not because its allowance is used up: Glider keeps serving in lean mode. It is offered under fair use, though. If an account sends traffic far beyond what Glider is meant for, GraphPilot may suspend its services and lock the account until the start of the next month, or until it moves to a paid plan. A warning email usually goes out before that happens, and nothing is deleted. Paid plans have no such limit; their usage above the allowance is billed.

What are request units?

Request units are what the request allowance and its overage are counted in. They are counted per request at each point of presence, and by response size. A request uses one request unit per started 10 KiB (10,240 bytes) of its uncompressed response body, at least one, at every point of presence it passes through. Headers are excluded, and the same query costs the same whatever the client's Accept-Encoding, so compression does not change the count. A response without a body uses one. Without shielding, a request passes through one point of presence: a 2 KB response is 1 request unit, a 25 KB response 3, and a 50 KB response, typical for a GraphQL API, 5. With shielding, a cache miss or a mutation goes on from the edge location that received it to the shield location you chose, so it passes through two, also when the shield takes too long to answer: a 2 KB response is 2 request units, a 25 KB response 6, a 50 KB response 10. A request the edge cannot send to the shield, and takes to your origin itself instead, passes through one. A background refresh of a stale copy through the shield is a request of its own. Purges through GraphPilot are free. Insights, and on paid plans its allowance of logged requests, count each logged request once, whatever its size or route; a service can log fewer requests or none (telemetry.record).