Quickstart
From a new account to a cached response in front of your own API, and a way to clear it.
GraphPilot documentation
GraphPilot is a managed edge cache and gateway that sits in front of your API, whether it speaks plain HTTP, GraphQL, or both. It caches responses at the edge on the rules your origin sends, lets you tag them and purge them by key, and protects your origin: it verifies credentials, limits what a request may carry, and signs every request it forwards to you. For a GraphQL API it goes further: it reads your schema, so a mutation clears the cached responses that contained the objects it returned, and it bounds what an operation may ask for. You describe it with one config file, plus your schema for GraphQL. There is no infrastructure to run.
Start here
Section titled “Start here”How GraphPilot works
Where GraphPilot sits, what it does for every request and what it adds for GraphQL, and what a service, a config, a schema and a deployment are.
Features and limitations
The feature matrix for plain HTTP and GraphQL, what is deliberately left to your origin, and the current limitations.
Reference
Every configuration setting, error code and response header, the policy expression grammar, and the CLI.
I want to…
Section titled “I want to…”- Cache signed-in data per organization, not per user: partitioning your cache
- Put GraphPilot in front of a REST API: caching a REST API
- Show a new item in a cached list right away: lists and creates
- Clear cached responses when my data changes: cache invalidation
- Find out why nothing is cached: why is nothing cached?
- Name my clients, retire old versions, refuse unknown callers: telling your clients apart
- Deploy from GitHub Actions or GitLab CI: deploy from CI
- Follow a request into my own logs: correlating with your logs
- Allow only the operations my clients send: migrating to trusted documents
- Test changes on staging, and roll back: staging and rollback
- Serve subscriptions over WebSockets safely: subscriptions and WebSockets
Every guide is listed on guides.
Get an account and the CLI
Section titled “Get an account and the CLI”Sign up at cockpit.graphpilot.io. If you do not belong to an organization yet, create one with New organization; the dialog asks you to accept the Terms of Service. The portal at cockpit.graphpilot.io is where you manage it. Then install the CLI and log in:
brew install graphpilot/tap/gpilotgpilot auth loginLinux, Windows and the other install methods are on installing the CLI. See signing in for device-code and API-key logins, and profiles for keeping several identities side by side.
Find your way around
Section titled “Find your way around”| If you are | Read |
|---|---|
| Evaluating GraphPilot | How it works, features and limitations, then limits |
| Integrating it for the first time | The GraphQL quickstart or the HTTP quickstart, then the connection to the origin and the go-live checklist |
| Moving from another CDN | Coming from another CDN |
| Using a specific server or router | Apollo Server, GraphQL Yoga, PostGraphile, federation routers, other servers |
| Caching data that differs per user or per organization | Who shares an entry |
| Clearing the cache when data changes | Invalidation |
| Working in the portal | Portal overview |
| Debugging a miss or an error | Troubleshooting, why is nothing cached?, headers, codes |
| Running it in production | Deployments, deploy from CI, staging and rollback, quotas |
| Reviewing it for security | Security and data protection, then hardening, authentication and request signing |
| Looking a term up | Glossary |
| Getting help | Troubleshooting, contacting support, and the status page |
| Reading offline or on paper | Download the handbook (PDF): every page of these docs in one document |