A technical audit is only valuable when its results lead to traceable decisions and verifiable improvements. This conviction has shaped the development of the Site Audit in Contextter from the very beginning. As the founder of an SEO platform, it was important to me not to stop at a list of detected errors. The real product value only emerges through a continuous workflow: define the scope of investigation, classify technical signals, identify root causes, derive actions, and then verify their impact.
We built the entire area along exactly this cycle. The current state connects configurable crawls, prioritizable findings, detailed URL and link data, segmented analyses, recurring audits, and controlled handoffs to AI agents.
The crawl begins with a clear question
Contextter can manage multiple websites within the Site Audit. We distinguish between your own domains and publicly accessible competitor websites. This distinction matters: for your own website, the focus is on fixing errors and technical improvement. For a competitor analysis, the goal is to examine structures and derive reliable insights for your own work.
Before starting, you can define which part of a website the crawl should cover. Available settings include the maximum number of URLs, crawl depth, include and exclude paths, and JavaScript rendering. This allows a team to examine exclusively a blog directory, a shop section, or another defined path without necessarily including the entire domain.
We have not yet published a binding maximum crawl size. We first need to standardize some internal parameters. At this point, a reliable product promise matters more to me than a number that might need to be qualified later.
After the crawl, classification comes first
The Overview condenses the most important results into an initial situation report. It shows the Health Score, the number of discovered and successfully crawled URLs, failed crawl attempts, and the distribution of issues across the severity levels Critical, High, Medium, Low, and Info.
These metrics are meant to provide quick orientation without preempting the analysis. If a significant proportion of discovered URLs could not be successfully checked, that is just as relevant as a small number of critical findings. Conversely, numerous low-priority hints may be less urgent than a single error affecting a commercially important page type.
The Health Score should be understood in the same way: it describes the technical condition within the audit but is not a promise of better rankings. Organic visibility depends not only on technical quality but equally on content, search intent, competition, and authority.
Issues become justified priorities
The issue views create the transition from overview to concrete work. Findings can be filtered by severity and reduced to the relevant subset. Severity provides an important technical assessment but cannot fully capture the business context.
A problem on five central landing pages can have a greater impact than a hundred informational hints in a barely relevant archive. That is why Contextter does not prescribe a supposedly objective to-do list. The system provides the technical evidence; the final order emerges from reach, page type, business goals, and the knowledge of the responsible team.
For me, this separation is essential. Good SEO software should support decisions but should not pretend that every prioritization can be automated without knowledge of the project.
Investigating root causes at the URL and segment level
Aggregated values show that a problem exists. To fix it, however, you need to see where it originates. That is why every crawled URL has its own detail view. It brings together HTTP status, canonical tag, robots directives, redirect information, and structured data, among other things.
The linking perspective adds another dimension: inbound and outbound links, the path through which the crawler discovered the URL, and metrics on internal, external, and nofollow-tagged links. This makes it possible to trace whether an existing page is barely reachable internally or was only found through an unnecessarily complicated path.
This level of detail also helps with root cause analysis. If the same finding repeats across many similarly structured URLs, it often points to a problem in the template or the underlying generation logic. Instead of correcting numerous individual pages separately, the team can address the shared root cause.
For larger websites, even the URL level is not always sufficient. A shop, an editorial magazine, and a help section can differ significantly in technical terms. The statistics view therefore allows analyses by directory, folder, locale, host, or URL pattern. HTTP status, click depth, indexability, discovery source, internal linking, crawl result, markup, and performance can all be analyzed.
The advantage of this segmentation lies in its precision. A good overall score can mask the fact that a single area is systematically flawed. Separate scopes make such deviations visible and also allow the development of a revised subsection to be observed independently from the rest of the domain.
Internal linking as a visible page network
Internal links connect navigation, discoverability, and content relationships. For this reason, we do not treat them as a subordinate metric but as a dedicated workspace.
The Link Explorer makes the captured relationships searchable, sortable, and filterable. Teams can check which pages receive the most internal references, which URLs those links originate from, and where recognizable gaps exist. Supplementary values for external and nofollow links complete the picture.
Beyond that, we are developing a view for Link Opportunities. It is intended to suggest possible internal connections. Editorial review remains necessary: the mere technical possibility does not justify a meaningful link. What matters is topical relevance, a helpful destination, and an integration that supports the reading flow. A sound internal structure is created through understandable relationships between pages, not through the highest possible number of references.
Tracking changes through crawl comparisons
An audit always describes the state at a specific point in time. Whether a migration, a release, or a round of corrections actually brought improvement can only be assessed through comparison over time.
The Compare function allows two completed crawls of the same site to be compared side by side. In addition to general metrics, Contextter shows changes in issues: which findings are new, which were resolved, and which have changed?
Interpreting these deltas still requires context. Declining issue counts can result from a successful fix but equally from a changed crawl scope or pages that are no longer reachable. A newly detected error does not necessarily have to have been caused by the last deployment. The comparison view provides the measurable change; for a reliable root cause assessment, the crawl configuration and development history must be taken into account.
Continuous monitoring instead of a one-time inventory
Websites change continuously: new content is added, templates are adjusted, redirects are modified, and technical releases are rolled out. Accordingly, Contextter supports recurring crawls. The associated schedules can be created, adjusted, paused, and restarted.
Alert rules complement this continuous monitoring. Recipients can be configured; due notifications are processed by a job and handed off to an email queue. This allows responsible individuals to react to relevant changes without having to wait for the next manual check.
In my view, alerts should be deliberately tiered. An overly broad configuration quickly generates too many notifications, while very narrow rules can miss gradual deterioration. Critical new problems are well suited for immediate notification. Smaller changes, on the other hand, can be bundled and evaluated in a regular technical review.
Controlled handoff to AI agents
In many teams, the analysis does not end with the responsible SEO specialist. Findings need to be handed off to development, editorial, or automated tools. For AI-supported workflows, Contextter therefore offers two different paths.
From a filtered issue view, a Markdown prompt can first be generated and copied. The selected findings become a structured briefing that can be passed to a coding agent, for example. This reduces manual preparation without taking away the team's control over the selected scope.
Additionally, an Agent Endpoint provides issues in Markdown or JSON format. Access is secured through Bearer authentication. On this basis, controlled workflows can be created in which an agent retrieves current findings and develops solution proposals.
The technical handoff does not replace expert review, however. Changes to canonicals, robots directives, or redirects in particular should be carefully reviewed before implementation. Errors in these areas can have far-reaching consequences, which is why automation here must be combined with clear checkpoints.
The current state of Contextter's Site Audit
Today, the area covers the essential stages of a repeatable technical SEO process: configurable crawls, a central Overview, filterable issues with severity levels, URL details, segmented statistics, link analysis, crawl comparisons, schedules, alerts, and handoffs to humans and agents.
In my view, this means the Site Audit has grown beyond the phase of a pure MVP. At the same time, there are points on which we cannot yet make definitive statements: specific limits, plan scopes, and an official launch date are still in progress.
The professional core, however, is established. Contextter is designed to map technical SEO as a closed improvement process: first, the scope is defined and crawled. Then results are classified, relevant issues prioritized, and root causes investigated at the URL, template, or segment level. From there, tasks are created for humans or agents. After implementation, another crawl comparison shows whether the measure achieved the desired effect.
As the founder, it is important to me to continue developing this area together with the requirements of professional SEO teams. Feedback on existing workflows and on the next meaningful expansion stages is therefore expressly welcome.