Sitebulb combines two products with different operating risks. Desktop runs audits on a Windows or Mac machine and keeps crawl data locally. Cloud runs on a provisioned server, shares results through a browser and scales beyond one operator's hardware. Both turn raw crawl observations into prioritized “Hints,” visualizations and reports.
> Distinctive strength: Raw crawl observations are translated into prioritized Hints, visualizations and communicable reports across Desktop or Cloud operation. > > Where it stops being an advantage: Severity labels can mislead unless the buyer checks raw responses, rendered pages and known-defect recall.
That presentation can make technical evidence easier to communicate. It can also tempt a buyer to accept severity labels without checking the raw response, rendered page and business context. The right Sitebulb test is not “Did it find many issues?” It is “Did it find the issues known to exist, avoid false alarms, explain the evidence and produce a repeatable decision?”
BenPicks has not run Sitebulb hands-on for this profile. The conclusions below come from current official product, pricing and support documentation. Vendor testimonials and comparisons are excluded from the verdict.
The Sitebulb decision in 60 seconds
| Question | Source-led answer |
|---|
| Best fit | Technical auditor who values explained findings and visual reporting; Cloud for shared team data and scale. |
| Poor fit | Buyer needing only raw exports, a content editor or an unreviewed issue score. |
| Trial | 14 days, full Pro-equivalent desktop features, no card. No Cloud free trial. |
| Desktop limits | Lite 10,000 URLs/audit and 100+ Hints; Pro 500,000 URLs and 300+ Hints. |
| Cloud limits | Starts at £95/month; up to 10 million URLs/audit depending on plan, with shared browser access. |
| Data boundary | Desktop audits are local; Cloud, Google integrations, Data Studio and MCP create separate flows. |
| Main unknowns | Exact live desktop price, Cloud assurance and deletion pack, binding SLA and complete exit export. |
Should you choose Sitebulb Lite, Pro or Cloud?
Lite is for occasional, smaller audits. Its 10,000-URL ceiling and 100-plus Hints can cover a small website, but it omits advanced configuration, comparisons, schedules and customized PDF reporting.
Pro raises the limit to 500,000 URLs per audit, exposes 300-plus Hints and adds the recurring-audit workflow. It still runs on the operator's machine, so available CPU, memory, disk and uptime matter. It is the practical default for an individual consultant or in-house technical SEO who audits regularly.
Cloud includes the desktop Pro licences associated with its user seats and moves crawl execution and shared evidence to a server. Public pages describe up to ten million URLs per audit, no project limits and no crawl-credit meter. Its strongest reason to buy is not just URL count: multiple people can work from the same audit without moving local databases or waiting for one laptop.
Do not choose Cloud merely because “unlimited” appears in marketing. Projects and audit counts may be unlimited while URLs per audit, server threads, concurrent runs, users, storage and plan price remain bounded.
How much does Sitebulb cost?
Cloud starts at £95/month and the desktop pricing page says large Cloud configurations can exceed £20,000/year. Final price changes with capacity and users. Monthly and yearly subscriptions renew automatically until cancelled; annual payment can save 15%, and invoice or purchase order is available for yearly plans by arrangement.
Desktop Lite and Pro prices are injected dynamically by currency, cadence and number of users. The accessible official page confirms GBP, USD and EUR, monthly/yearly choices and additional user pricing, but did not expose a durable static amount suitable for this profile. That is a genuine pricing uncertainty, not a reason to copy a stale cached number. Capture the live checkout on the decision date with currency, tax, users and renewal cadence.
Total cost includes operator time to validate Hints, local hardware for Desktop, Cloud capacity and any reporting or integration work. A cheaper crawler is not cheaper if reviewers spend days dismissing noisy findings.
Is the free trial useful?
Every user gets a fully featured 14-day desktop trial equivalent to Pro, without a card. That is enough to build a representative project, compare HTML and JavaScript crawling, test advanced settings and export a real audit.
Cloud has no free trial. The vendor suggests using desktop because feature behavior is similar, then requesting a Cloud demonstration. That cannot prove team collaboration, server throughput, access control or exit. Put those items into a demo script and written order rather than assuming desktop parity.
What are Sitebulb Hints—and can you trust them?
Sitebulb analyzes crawl data into Hints with severity, explanation and affected URLs. Categories include on-page SEO, resources, indexability, structured data, internationalization, performance and security. Some expensive checks, such as similar-content analysis, can be off by default.
A Hint is a documented rule applied to observed data. It is not proof that the issue harms traffic, that every affected URL should change or that the suggested priority matches your business. A canonical difference may be intentional. A blocked page may be correctly blocked. A duplicate title across parameter variants may require URL control rather than copy changes.
For a pilot, sample at least 30 Hints across severity and category. Check raw HTTP, source HTML, rendered HTML, robots rules, templates and the intended site architecture. Record true issue, intentional condition, crawler limitation, stale evidence or false positive. Only accepted findings count toward workflow value.
HTML crawler or Chrome crawler?
The HTML crawler is faster and shows what the initial response exposes. The Chrome crawler renders client-side JavaScript and can find content or signals added later. Sitebulb itself recommends a sample crawl and comparison of response versus rendered HTML before deciding.
The choice changes runtime and findings. HTML speed uses thread and maximum-URL-per-second controls. Chrome mode uses renderer instances and render timeout as well as a URL cap. Increasing concurrency can load the target server and the crawler machine.
Run the same bounded URLs through both modes. Compare links, canonicals, robots, content, structured data, status and resources. Explain each difference. Do not treat one mode as universally “more accurate”; it answers a different representation question.
What changes between Desktop and Cloud data?
Desktop stores projects and audits on the installed computer. Changing the licence does not delete those local projects as long as the same machine remains available. This is simple, but backup, disk failure, device encryption and hand-off are your responsibility.
Cloud stores audits in a workspace on a provisioned tenant server. Admins manage user seats, secret keys, concurrency, queue policy, archive schedules, webhooks, Slack and Data Studio access. Concurrent crawling is reserved for Large and Custom plans; otherwise work queues when threads are unavailable.
The documentation warns that the admin secret key is tied to the server and cannot be deleted. Treat it as a high-value credential: never share it, use separate revocable keys for users and integrations, and obtain the incident and rotation process before purchase.
Does Sitebulb have an API or MCP integration?
The Cloud FAQ says there is no general API. Webhooks and Data Studio expose aggregated or dashboard data. For complete affected URL lists, use application exports.
MCP now provides 15 read-only tools. It cannot run, pause or delete crawls, change settings or write data. Authentication is per-user OAuth with a single read scope over TLS, and connections are individually revocable. Nothing syncs in the background; data moves only when a question requests it.
The important boundary begins after the answer reaches the AI assistant. From that point, the assistant provider's agreement and settings govern the data. Connect only a non-sensitive fixture first, confirm the chosen workspace or desktop machine and test revocation. Do not paste OAuth codes into chat.
Which integrations increase risk?
GA, GSC, Google Sheets and Data Studio can enrich or export audit data. Enable them per project where possible rather than globally. Record account, scopes, retention, owner and revocation. A crawler may observe unpublished or authenticated material; sharing its export can reveal URLs, content, headers, security findings and business performance.
Cloud's collaboration benefit therefore needs role testing. Create the smallest representative user set, verify what each role can see, remove one user, revoke a secret key and confirm access disappears. Browser convenience should not widen access silently.
Who should shortlist Sitebulb?
Choose Desktop Pro when one technical operator wants stronger explanations and reports than a raw crawler, can maintain the machine and needs repeat audits under 500,000 URLs. Choose Cloud when shared evidence, remote collaboration, scheduled server crawls or higher scale justify materially higher spend.
Look elsewhere when the requirement is only a lightweight small-site crawl, a fully scriptable general API, or centralized security evidence that Sitebulb cannot provide during procurement. Compare it with other options in AI SEO software, the technical SEO crawler comparison guide, and the enterprise AI SEO buying guide.
A fail-closed Sitebulb evaluation
- Create a fixture with known redirects, canonicals, robots rules, duplicate content, hreflang, structured data, JavaScript-only links and security headers.
- Record expected findings and affected URLs before running the tool.
- Use the Pro trial to run an HTML crawl with explicit URL, depth, thread and rate limits.
- Repeat the same scope with Chrome and explain response-versus-render differences.
- Validate at least 30 Hints against raw evidence; count accepted findings, intentional conditions and false positives.
- Change five known properties, rerun and verify audit comparison finds those changes without overwhelming noise.
- Measure crawl time, requests, CPU, memory and disk; project the production workload.
- Export complete affected-URL lists, project settings and a stakeholder report; verify another reviewer can reproduce conclusions.
- If considering Cloud, test queueing, collaboration, roles, user removal, revocable secret keys and archive behavior in a demo.
- Connect one required Google or MCP integration using non-sensitive data, then revoke it and verify the flow stops.
- Request Cloud assurance, regions, encryption, backup deletion, incident, uptime, support and exit-export terms.
- Compare annual total cost with accepted issue decisions and reporting time—not raw Hints generated.
Pass only when the crawler matches fixture truth, settings are reproducible, target load is safe, reviewers can explain priority, data flows are approved and exit evidence is portable. A polished PDF cannot compensate for an unverified Hint.
Official sources