# Graphwake

> Graphwake maps which teams depend on each endpoint of your OpenAPI specs, shows the impact of a change before it merges, and tells only the teams it affects.

## Who it is for

Engineering teams that run many services which call each other's APIs, and the engineering leaders responsible for them.

## What makes it different

Change detection tools tell you what changed in an API. Graphwake also knows who is affected: consumer teams subscribe to the endpoints they use, and each breaking change is routed only to those teams, through Slack, Microsoft Teams or email.

## Status

Available now. Every workspace starts with a 7-day trial of every feature, with no credit card. Plans are a flat monthly price set by the number of watched endpoints, plus a monthly fee for each user beyond the plan’s included users: Basic, Pro and Enterprise. Prices are on the /pricing page.

## FAQ

### What is Graphwake?

Graphwake is a dependency map for your APIs. It watches your OpenAPI specs, knows which teams consume each endpoint and which team owns it, and works out who a change affects before it ships. Only the people affected by a change hear about it, and every change and decision is kept in one history.

### Who owns an endpoint?

Any team can be named the owner of a service or of a single endpoint. Ownership is optional: endpoints without an owner, or without any consumer, are flagged so someone can claim or retire them.

### How does Graphwake know who a new endpoint affects?

When an endpoint is added or changed and no team is mapped to it, Discovery compares it with the rest of your specs and subscriptions: teams that use neighbouring endpoints, endpoints that share the same schema, who owns what, and suggestions teams accepted or dismissed before. Each suggestion comes with a score and the reason behind it, from clear rules rather than a language model. A team becomes a subscriber only when someone confirms it.

### How is Graphwake different from Bump.sh or oasdiff?

oasdiff compares two OpenAPI files and finds breaking changes. Bump.sh publishes API docs and changelogs. Both answer "what changed?". Graphwake also answers "who is affected, and how badly?": consumer teams subscribe to endpoints, each pull request gets an impact score before it merges, and each breaking change is routed only to the teams that use that endpoint.

### What counts as a breaking change?

A change that can make an existing client fail: removing an endpoint or a response field, adding a required parameter or request field, changing a field's type, or narrowing the values a field accepts. Additions such as a new optional field are not breaking and stay quiet by default.

### Do you need access to our code?

No. Graphwake only needs your OpenAPI spec. You either give us the URL where the spec is published, or a step in your CI pipeline sends the spec to us.

### Where do alerts show up?

In Slack, Microsoft Teams or email. Each team chooses its channel, and an alert names the endpoint, what changed, who changed it and which pull request did it.

### Which CI systems are supported?

The CI agent works with GitHub Actions, GitLab CI, Azure DevOps, Bitbucket Pipelines and Jenkins. It never fails or blocks your pipeline. If you’d rather not touch CI at all, give us your spec’s URL and we check it every 30 minutes.

### How does pricing work?

A flat monthly price set by the number of endpoints you watch, plus $1.99 a month for each person beyond the plan’s included users. Basic is $49 a month for up to 150 endpoints and 10 users, Pro is $119 a month for up to 500 endpoints and 30 users, and Enterprise is priced with you.

### What happens when the trial ends?

The trial lasts 7 days and includes every feature. You don’t enter a card to start. Before it ends, pick a plan from the billing page; if you don’t, the workspace closes and nobody is charged.

### What if we go over our plan’s limits?

Nothing stops working. Alerts keep going out and every endpoint stays watched; Graphwake tells your admin it’s time to move up a plan.

### How do I get started?

Sign up with your work email, paste a spec URL or add the CI step, and invite your teams. Each team marks the endpoints it calls, and the next breaking change reaches them. Most of that fits in an afternoon.