Build visibility people and AI can find
Learn how to audit a website, measure visibility across search and AI answers, turn evidence into action, and share the result.
Sight brings technical search auditing and AI-answer visibility into one evidence workspace. A project starts with a real website, a defined brand profile, and a tracked set of audience questions. Each scan then preserves the crawl results, model responses, citations, competitors, and configuration needed to explain how a conclusion was reached.
Use these docs as an operating guide rather than a feature catalog. The sections explain how to prepare a project, interpret SEO and GEO measurements, inspect the evidence behind an issue, work with sources and actions, and communicate findings in a report. Limits and availability are documented separately so observed behavior is not confused with planned capability.
A typical review moves from scope to evidence to action. Confirm what the scan covered, identify the strongest changes or gaps, open the supporting URLs and model responses, and only then decide what deserves work. When a finding is shared outside the workspace, include enough context for another person to reproduce the view and understand what Sight did and did not measure.
Run your first scan
Create a project, complete the brand profile, select prompts and models, and start a combined SEO and GEO scan.
Understand the workspace
See how projects, scans, prompts, answers, sources, actions, and reports fit together.
Connect your data
Use Google Search Console, webhooks, the Pro REST API, and report sharing.
Built for people and agents
Connect Sight Docs to AI clients and developer workflows through our supported MCP server, dependency-free CLI, or installable agent skill.
MCP server
Connect compatible AI clients to read-only tools for listing, searching, and retrieving public Sight documentation.
Docs CLI
Search the complete documentation corpus from a terminal and return Markdown or structured JSON.
Agent skill
Install a compact skill that teaches agents how to verify availability, limits, and canonical sources.
One workspace for SEO and GEO
Sight keeps classic website evidence and AI answer evidence separate, then brings both into one project.
SEO and GEO remain separate measurement systems because they observe different surfaces, but they share project identity, dates, filters, and reporting. Keeping the evidence distinct prevents a technical crawl finding from being mistaken for an AI-answer observation while still allowing both to support one prioritization decision.
| Workspace | What it answers | Evidence |
|---|---|---|
| SEO | Can search engines crawl, understand, and rank this website? | Crawled pages, rendered pages, metadata, links, structured data, performance, robots and sitemap checks |
| GEO | Does the brand appear in AI answers, and which sources shape those answers? | Tracked prompts, model responses, citations, competitors, sentiment, position, fanouts, ads and source gaps |
| Reports | How do we turn findings into a client-ready narrative? | Editable report blocks, presentation and document layouts, share links, PDF export and delivery history |
Documentation scope
These docs describe behavior that exists in the current Webreport codebase. Planned capabilities are marked clearly on Feature availability.
When the code and docs disagree, treat the product behavior as the issue to investigate and update this documentation only after the intended contract is confirmed.