Intercrawl and self-hosted prerendering
Running your own prerender server or Rendertron, or using a service. What you run yourself, what you get, and what Intercrawl adds, from each project's own pages.
Last checked 21 October 2026
Some sites serve crawlers a rendered copy from a server they run themselves: the open source Prerender server (prerender on npm), or Google's Rendertron. Intercrawl does the same job as a service. This page sets out what each asks of you. What it says about the two projects comes from their own documentation and source, listed under Sources with the date each was read.
At a glance
| Prerender server | Rendertron | Intercrawl | |
|---|---|---|---|
| What it is | A Node.js server that "uses Headless Chrome to render HTML" | A rendering server from the Google Chrome team | A hosted service |
| Maintained | Version 5.21.6, published 2024-09-12 | Deprecated; the repository was archived on 2022-10-06 | Yes |
| Licence | MIT | Apache 2.0 | A hosted service; its install files are yours to use on your own sites (Terms) |
| What you run | The server and a Chrome install ("Prerender does not download Chrome") | The server and its headless Chrome | Nothing: one install file, a connection, or a DNS record |
| Cache | Through plugins, such as an in-memory cache | A built-in cache, with refreshCache=true and /invalidate/<url> | 7 days, renewed when the page changes |
| Middleware in front of your site | prerender-node and the other Prerender middlewares | rendertron-middleware for Express (version 0.1.5, 2018) | Intercrawl's install files, or Prerender's middlewares |
| How a crawler is recognised | By the middleware, from the user agent | By the middleware, from the user agent | By the user agent, and the IP address against the operator's published ranges where they exist |
| Cost | Your server and its upkeep | Your server and its upkeep | Free during the 14-day trial, and free after it until paid plans open |
Running your own server
The Prerender server is a Node.js project: you install it with npm, start it, and point a Prerender middleware at it with PRERENDER_SERVICE_URL. Chrome must already be installed on the server. Caching, access control (a basic auth plugin) and a list of allowed domains are plugins you turn on in server.js. You keep the server up, sized for your crawler traffic, and patched.
Rendertron works the same way, with GET /render/<url> as its API. Its README says "this project is deprecated" and "Rendertron will not be actively maintained at this point", and its GitHub repository has been read-only since 6 October 2022. Its Express middleware sends crawlers' requests to the server with no token, and its default list of crawler user agents (16 entries) names Baiduspider, Bingbot and social and link-preview crawlers, and none of Googlebot, GPTBot or ClaudeBot.
What a service changes
With Intercrawl there is no server or browser to run. The render API is compatible with Prerender's middlewares, so a site that uses prerender-node with its own server can switch by changing the service address and adding a token. Beyond that:
- Crawler checks. Where a crawler's operator publishes the IP addresses it crawls from (Google, Bing, OpenAI, Anthropic, Perplexity and Apple do), each request is checked against them, and a request that names Googlebot from any other address gets your own page. See How serving works.
- A time limit. A crawler waits at most 10 seconds; after that, or on any error on Intercrawl's side, your own page is served.
- Renewing the cache. A page is rendered again after 7 days, when its scripts, stylesheets, title, description or canonical change, when your sitemap gives it a newer
lastmod, and when you call the re-cache API. - Checks of what crawlers received. A daily audit of up to 100 pages compares what Googlebot, GPTBot and ClaudeBot got with what a visitor got, with alerts by email, Slack or webhook. See Audits and alerts.
What you give up: the renders are made on Intercrawl's browsers, within its limits (1,000 fresh renders per site per day), and a self-hosted server can be shaped in code in ways a service cannot.
Moving to Intercrawl
Move from self-hosted prerender or Rendertron has the steps: for the Prerender server, two settings in your middleware; for Rendertron, replacing its middleware with an Intercrawl install.
A note on rendering for crawlers
Google calls serving crawlers a rendered copy (dynamic rendering) "a workaround and not a recommended solution", and recommends "server-side rendering, static rendering, or hydration" instead. It also says that "Googlebot generally doesn't consider dynamic rendering as cloaking" as long as it "produces similar content". Every option on this page renders for crawlers. If your site can render on the server, that is the better fix; Intercrawl's audits are there to show the crawler's copy matches what visitors get.
Intercrawl does not promise any ranking, indexing or citation by an AI answer, and neither does this comparison.
Sources
Read on the dates shown (UTC). Projects change; if anything here is out of date, write to hello@intercrawl.com.
- prerender on npm, version 5.21.6 and its README, read 2026-10-21
- Rendertron on GitHub, read 2026-10-21
- rendertron-middleware source, read 2026-10-21
- Google Search Central: dynamic rendering as a workaround, read 2026-10-21