Blog · How to use CF Ready
How to Run Your First CF Ready Scan
Run npx @orangecloud/cf-ready scan or use the Web Agent, then read blockers, apply only the safe fixes, and optionally repeat the scan in CI.
Run a scan without a long setup
From the project you want to launch, run npx @orangecloud/cf-ready scan. The cf-ready scan command inspects the repository on your machine. It detects the framework, package manager, routes, config files, and likely deployment target, then scores migration, security, AI readiness, SEO, and deployment. It writes a readiness report, including cf-ready-report.md, cf-ready-report.json, migration-plan.md, and go-live-checklist.md. Nothing is deployed.
If you would rather not install the CLI, use the Web Agent at https://ready.orangecloud.vn/app/. Upload a ZIP of the source, or import a public GitHub repository, then run scan in the browser terminal. Private repositories need GitHub sign-in on the agent. You get the same kind of readiness report either way, which is useful when a reviewer does not have the repo cloned locally.
Read the scores before the prose
Start with the overall number and the category breakdown, then read the blocker list. A filesystem write in a route handler, a committed .env file, or a missing build script will matter more on launch day than a missing Open Graph image. Severity runs from blocker and high down through medium, low, info, and passed. Any blocker means the project should not be treated as production-ready yet.
The narrative files explain the same findings in sentences you can paste into a ticket. migration-plan.md is where a recommended Cloudflare path shows up, including notes for Next.js projects that may belong on Workers through a supported adapter. The plan is advice. The tool does not run the migration, does not call vinext init, and does not change routing or auth code on your behalf.
Fix only the safe gaps
When the report offers an automatic fix, run the narrow command you actually want. cf-ready fix --seo can draft a sitemap and metadata suggestions. cf-ready fix --ai-readiness can draft robots.txt, llms.txt, llms-full.txt, an OpenAPI file, and an MCP server card. Open the diff and read it. The files are starting points for a readiness report you stand behind, not a claim that the site is finished.
Skip the fix commands if you only wanted the assessment. A scan is complete without them. Adding files you have not reviewed is how a helpful draft becomes a wrong canonical URL or a robots rule that blocks the wrong path.
Repeat the scan on pull requests
The CF Ready GitHub Action runs cf-ready scan in CI, uploads SARIF, and can fail the job when blockers exist. Pin the action to a released tag from the sycu8/Cloudflare-Go-Live-Readiness repository, and point it at the directory that contains the app. A companion action can open a pull request that contains only the safe AI and SEO assets. It does not merge that pull request. Branch protection stays in charge of what lands.
CI is also where the readiness report stays honest. Local scans drift when someone forgets to re-run them after a dependency bump. A pull request check makes the score part of review, next to tests.
After the first green report
Re-run npx @orangecloud/cf-ready scan when the dependency tree or the framework config changes. Before you announce the launch, run cf-ready deploy-check for scripts and Wrangler config, then cf-ready smoke-test with your production URL. The smoke test confirms the homepage, robots.txt, and sitemap.xml respond, along with redirect behavior and basic security headers.
Command docs live at https://ready.orangecloud.vn/docs/commands.html and the Web Agent is at https://ready.orangecloud.vn/app/. The first scan is the baseline. The smoke test is the proof that the deployed site matches the readiness report you signed off.