By the iSuggest.ai Team · Updated for 2026
Type a URL into the home page, click "Run free audit," and a report appears. No spinner that says "Step 2 of 6," no email that arrives twenty minutes later saying your results are ready. It's easy to assume something is quietly running in the background and the page is just polling until it's done. It isn't. It's worth explaining what actually happens between the click and the report, because the honest answer is both simpler and more constrained than most people guess.
One request, one response
When you submit a URL, iSuggest.ai makes a single HTTP request to the audit engine and waits for it to answer. There's no job queue, no background worker picking up your request later, no polling loop checking "is it done yet?" every few seconds. The request goes out, the engine fetches your page, analyzes it, and sends back a JSON payload of findings. When that response lands, we store it and redirect you straight to /reports/{id}. The whole thing is one round trip.
That matters because it rules out a whole category of things that can go wrong with queued systems: a job stuck behind someone else's backlog, a webhook that never fires, a report that silently vanishes because a worker crashed. If our audit engine answers, you get a report. If it doesn't, you get an error immediately, not a silent queue entry.
What the engine actually does in that one request
A routine audit calls the engine's single-page endpoint with your URL and asks it to respect the site's own robots.txt before fetching anything. The engine loads the page the way a crawler would, then runs it through the checks described in what we actually measure: page identity and meta tags, heading structure, internal and external links, image alt text, structured data types, crawl status and response time, and a word-count/text-ratio read on the content itself. All of that comes back in one response — there's no second call to fill in missing pieces for a routine audit.
The one exception is publishing. If you later submit that page to an AI directory, a separate, deeper capture runs at that moment — full readable content, headings, JSON-LD entities, links with anchor text — because that snapshot has to stand on its own once it's frozen. A routine audit doesn't do any of that extra work, which is part of why it stays quick and why it's free.
Why it isn't instant
"Fast" doesn't mean zero time. The engine has to actually fetch your page over the network, and that step is entirely at the mercy of your server: a page that responds in 200ms comes back quickly, a page sitting behind a slow origin server or heavy render pipeline takes longer, the same way it would for any bot or human visitor. We don't fake a faster result by sampling a cached copy — the audit reflects the page as it responds right now, which is also why a slow time-to-first-byte shows up as its own finding rather than just a background irritation.
On our side, the request has a hard ceiling: if the engine hasn't answered within 90 seconds, the call times out rather than hanging indefinitely. In practice that ceiling exists for pages that are badly broken or unreachable, not for a normal page load — but it's a real limit, not a marketing number, and we'd rather tell you it exists than pretend every audit finishes instantly.
The one-a-minute limit, and why it isn't about money
Signed-in accounts can run one audit per minute. It's tempting to assume a rate limit exists to nudge you toward a paid tier, so it's worth being direct about why this one exists: our audit engine calls out to services that have their own rate limits upstream, and a shared limit at our end keeps one account from accidentally (or deliberately) hammering those services for everyone else. It has nothing to do with credits — as covered in how the credit system actually works, running audits is free and unlimited in the sense that there's no cap on how many pages you can check over time. The one-per-minute rule is a pacing limit, not a usage cap, and it resets on its own; there's no support ticket required to get your next audit.
What "free" does and doesn't include
Because Cluster 6 of our own content guardrails exists precisely to stop us from overselling this: audits are free, but "free" describes the audit, not everything you can do with the result. Reading your report, seeing every issue with its evidence and suggested fix, and re-running the audit as often as the pacing limit allows — all of that costs nothing and requires no account. What costs a credit is the separate act of publishing a frozen snapshot of a page to a public AI directory listing, because that step runs the deeper capture described above and stores something permanent. The audit itself was never the paid part, and it still isn't.
What this means if you're deciding whether to run one
The practical takeaway is that there's very little reason to hesitate before running an audit. It's one request, it's free, it doesn't queue you behind other users, and the only real constraint is a one-minute pacing gap between runs on the same account. If you want to see the plain mechanics of what a report is built from before you click, how it works walks through the same flow end to end. And if the report tells you something worth fixing, you don't need to wait for a scheduled re-check — the next audit is one minute away, not tomorrow.