By the end, your published app will rank higher in search engines — and the SEO Score in the Publishing tool will reflect it. After every successful publish, Replit runs an SEO audit and shows a 0–100 score. If the score has room to grow, one click hands the specific issues to Agent. For most builders, the workflow is just: publish → glance at the score → ask Agent to optimize. For details on how the score is calculated (Lighthouse audit, thresholds, supported deployment types), see SEO Score.Documentation Index
Fetch the complete documentation index at: https://docs.replit.com/llms.txt
Use this file to discover all available pages before exploring further.
Publish your app first
You won’t see an SEO Score until your app has been published. If you haven’t yet, publish from the Agent chat or the Publishing tool, then come back here.Optimize with Agent
When your SEO Score is below 90, an Optimize SEO with Agent button appears next to the score in the Publishing tool. Selecting it sends the exact action items Lighthouse identified — for every URL it checked — straight to a fresh Agent chat with instructions to fix them. After Agent finishes its changes, republish your app. The SEO Score recalculates against the new build and the badge updates automatically. Common fixes Agent will apply:- Add or rewrite missing or generic
<title>tags so each page has a unique, descriptive title. - Add a meta description to every page.
- Add descriptive
alttext to images. - Set the
langattribute on the<html>element. - Make sure links have discernible text (no “click here” or empty anchors).
- Add a
robots.txtand asitemap.xml. - Add a viewport meta tag for mobile rendering.
- Improve color contrast on text that fails accessibility checks.
- Add Open Graph and Twitter card meta tags so links to your app preview well on social platforms.
Best practices for an SEO-optimized app
Whether you ask Agent to handle these for you or write them yourself, these are the highest-leverage things to get right:- Give every page a unique title and meta description. Aim for titles under 60 characters and descriptions around 150–160 characters.
- Use semantic HTML. Wrap your layout in
<main>,<header>,<nav>, and<footer>. Use<h1>once per page and step heading levels in order. - Add alt text to every image. Describe what the image shows, not just what it is.
- Generate a
sitemap.xmlandrobots.txt. Ask Agent to create both so search engines know what to crawl. - Add Open Graph and Twitter card tags. Control how your app looks when shared on social media.
- Use a custom domain. A branded domain builds trust with both visitors and search engines. See Add a custom domain.
- Pick the right deployment type. For content-heavy sites, Static Deployments deliver pre-rendered HTML that search engines parse instantly.
- Add structured data. JSON-LD markup helps search engines show rich results for products, articles, FAQs, and events.
- Keep your app fast. Optimize images, avoid render-blocking scripts, and lean on Replit’s hosting to keep response times low.
Improve it next
SEO Score
See exactly how the score is calculated, the threshold ranges, and which deployment types support it.
Add a custom domain
Move your app to your own domain — a branded URL builds trust with both visitors and search engines.
Static Deployments
The best deployment type for content-heavy, SEO-critical sites.
Publishing Geography
Choose where your app runs to keep response times low for your audience.
Need further help?
- The score didn’t update after a republish: wait a minute or two — the audit runs asynchronously after the deployment goes live, and the badge refreshes when results land.
- The score is the same after Agent’s changes: ask Agent which specific Lighthouse items it addressed, then check the badge tooltip for the items that are still flagged.
- Some pages have a lower score than others: the badge surfaces the lowest score across all crawled URLs. Ask Agent to focus on the specific page Lighthouse flagged.
- You don’t see a score on a published app: SEO Scores only run for public web-facing deployment types (Autoscale, Reserved VM, Static). Scheduled and private deployments don’t produce a score.