Methodology
How we score
We check how developers find your product, get started, and get help. Every score runs from 10 to 95, calculated by code from what we can measure.
How the overall score works
Your DevRel Score is a number from 10 to 95. It answers one question: how easy is it for a developer to find your product, start using it and stick with it? Every score we show, overall, per link and per area, stays in that range: there is always something to improve, and anything we could measure counts for something.
We look at the places developers actually meet you: your docs, your GitHub, X, your Discord community, other channels like YouTube, and what AI models say when developers ask them for advice. Each link you give us gets its own score, and together they make up your overall score.
Score ranges
Every score runs from 10 to 95- 80–95Strong
- 50–79Good, with gaps
- 10–49Needs work
What we look at
DocsWhether a new developer can follow your quickstart to a working result, and whether it is a real getting-started guide rather than a copy of your README.
Whether a new developer can follow your quickstart to a working result, and whether it is a real getting-started guide rather than a copy of your README. Whether the docs are organised so people can find their way: getting started, guides, concepts and reference, each in its place, without the same pages repeated in two sections. Whether your pages show real code they can copy, whether links lead where they should, and whether the docs are kept up to date. We also look at how easy your docs are for AI tools to read, because more and more developers meet your docs through an AI assistant before they ever open a browser tab, but that alone can't make messy docs score well.
GitHubHow quickly you respond to the people who open issues and pull requests, how easy it is for someone new to make their first contribution, and how clearly your README explains what the project is, how to install it and where to go next.
How quickly you respond to the people who open issues and pull requests, how easy it is for someone new to make their first contribution, and how clearly your README explains what the project is, how to install it and where to go next. We also look for signs that the project is alive and safe to depend on.
Other channels (YouTube today)Whether you publish regularly, whether your videos teach people how to build things with your product, and whether viewers can easily get from a video to your docs or code.
Whether you publish regularly, whether your videos teach people how to build things with your product, and whether viewers can easily get from a video to your docs or code. We also look at whether your videos are still finding an audience.
AI VisibilityDevelopers increasingly ask an AI model "what should I use for…?" before they search.
Developers increasingly ask an AI model "what should I use for…?" before they search. So we do the same: we ask several AI models the kinds of questions developers ask in your category, and see whether they suggest your product. Your report shows how often that happened, for example "Suggested in 9 of 20 AI answers".
We also look at how the models talk about you: whether they describe your product correctly, whether they point people to your own docs, and how often they name you compared with the alternatives.
AI Visibility only needs your product name and category, so it runs even if you don't give us any links.
XWhen we have your X handle (you gave it, or your profile links your site), we read your account's recent public posts, replies and mentions with the official X API: how regularly you post, how people engage, whether you answer the people who mention you, whether posts link to your docs, repository or videos, and the mix of formats.
When we have your X handle (you gave it, or your profile links your site), we read your account's recent public posts, replies and mentions with the official X API: how regularly you post, how people engage, whether you answer the people who mention you, whether posts link to your docs, repository or videos, and the mix of formats. Every audit includes X, with or without a login.
DiscordAdd our bot to your server and we read the channels it can see: how fast questions get a first reply, how many go unanswered, whether your team or the community answers, and the questions that keep coming back, which usually point at a gap in your docs.
Add our bot to your server and we read the channels it can see: how fast questions get a first reply, how many go unanswered, whether your team or the community answers, and the questions that keep coming back, which usually point at a gap in your docs.
Where the data comes from
Everything in an audit comes from public sources: the public pages of your docs site, the public GitHub and YouTube APIs, your X account's public posts and mentions through the official X API, and answers from AI models to general questions about your category. We don't need access to anything private, and we never ask for it.
Discord is the one source you connect yourself: add our bot, and we only read the channels it can see. We'll only ever use what you choose to connect.
Our principles
- Scores are calculated by code, not by an AI. Every score comes from fixed rules applied to the data we collected.
- The same data always gives the same score. We keep the data behind every report, so the same results are always scored the same way.
- AI writes suggestions, not scores. We use AI only to help write the improvement suggestions in your report, and every suggestion has to point at something real we found on your links.
- You can see what was measured. Your report shows which links were scored and which weren't, so anyone looking at your score knows what it covers.
What your score means
| Score | What it means |
|---|---|
| 80 to 95 | Strong. Developers can find you, get started and get help without friction. Keep it that way. |
| 50 to 79 | Good, with gaps. The basics are there, but some things are slowing developers down. Your report shows which ones. |
| 10 to 49 | Needs work. Developers are likely getting stuck or not finding you at all. The good news: the first few fixes usually make the biggest difference. |
FAQ
Why did my score change?
Your score is based on live data: your docs, your repository, your videos and what AI models are saying right now. When those change, your score changes too. New issues, a new release, a fixed broken link or a different answer from an AI model can all move the number. AI answers in particular vary from day to day, so small movements in AI Visibility are normal.
Why is one of my links "not measured"?
Either you didn't give us that link, or we couldn't read it during the scan (for example the site was down, or the repository or channel is private). Your overall score only reflects what we could measure, and the report says so. Run a new scan once the link is reachable.
How do I improve my score?
Improve the things developers actually notice. Every report is free and needs no login: it shows every check we ran, whether it passed, and what we found, so you can see what is holding each link back. Fix those, then run a new audit to see the difference.
Signing in is optional: it saves your reports, and each new audit of the same links shows what changed since the last one, so you can re-scan after a change and watch your score move. Pro ($5 one-time for a report's fixes) unlocks every fix, with step-by-step instructions for your real pages, ready-to-paste drafts and what to fix first. Pro is coming soon. Run a scan to see where you stand.
Can I see exactly how each score is calculated?
No. We don't publish the individual rules, thresholds or weights. If we did, it would be easy to raise the number without making anything better for developers, and the score would stop meaning anything. What we do promise is that the same data always gets the same score, and that every improvement we suggest points at something real on your links.
See where you stand
Paste a link and get your DevRel Score, every check and what we found in a few minutes. Free, no login needed.