What is EdgeComet?
EdgeComet is an infrastructure platform for SEO. It sits in the request path between your website and the bots that crawl it. Google and AI crawlers receive complete pages. Everything they receive is recorded, and that record drives the analytics and automation for improving how the site performs in search.
Setup requires no code changes on your site: bot traffic is routed through EdgeComet at your CDN or proxy, human visitors keep hitting your site directly. See Getting Started.
What you get
Caching
EdgeComet takes bot traffic off your servers.
Every bot request is answered from EdgeComet's HTML cache. Your servers build a page only when the cache refreshes it, not once for every bot that asks. What still reaches them is a small fraction of today's bot load, and the Search Console Crawl Stats report shows load time under 100ms.
JavaScript rendering
Pages are rendered in headless Chrome. Crawlers that do not execute JavaScript, including the major AI bots, receive complete content.
The renderer is built for the hard cases: it waits for the page's own late requests to finish and scrolls the page to the bottom, capturing sections that mount only on scroll. A render that fails or times out falls back to the last good cached copy instead of a broken page.
Bot analytics
Per-bot crawl behavior from real in-path traffic, not from a sample.
Which bots fetch which pages, how often, what they received, and how much of your crawl budget is wasted. AI bots are tracked individually and split by purpose: training, index refresh, and live user-triggered fetches.
Continuous site audit
EdgeComet runs no artificial crawl of your site. You see it as Googlebot, Bingbot, and the AI bots see it.
Indexation, on-page tags, canonicals, hreflang, internal linking, and technical health, with every metric computed from the pages bots actually received. The audit updates the moment a bot fetches a page, with no scan schedule and no extra crawl load.
Search analytics
EdgeComet pulls the full Search Console dataset through the API and joins it with its own crawl and log data, giving every page the impressions, clicks, and positions it earned.
Keyword portfolio, traffic decay, cannibalization, and section-level performance. The same data carries the AI-side analysis, where fan-out queries are identified and clustered back into the short user prompts behind them, and it ranks every issue the platform reports by the traffic at stake.
Edge SEO
Change titles, meta descriptions, canonicals, hreflang, and any element matched by a CSS selector, insert internal links, and set redirects and status codes.
Rules apply to the rendered DOM and are baked into the cache, with no origin code change and no release to wait for. AI agents draft them for you, either EdgeComet's own Action Board or your assistant over MCP, and every agent-authored change lands as a draft a human reviews and deploys.
Action Board
The centerpiece of the platform, working across every data layer at once: bot renders, crawl logs, and Search Console joined on the same pages.
It is an AI agent that constantly watches your website, spots issues, analyzes the data behind them, and prioritizes everything by the traffic impressions and AI impressions at stake. From there it drafts the fix and stages it in Edge SEO. The same engine is available conversationally: ask questions about your site in plain language and get real data back.
Alerts
Anomaly detection runs on live bot traffic, not on a scheduled recrawl of your site.
Rules evaluate every five minutes, and a broken page, an indexability regression, or a wave of 5xx responses reaches you within minutes of the bots meeting it. An outage does not stop the crawl either: EdgeComet keeps serving the cached pages, and Googlebot and the AI bots go on reading your site while the origin is down.
Self-configuring setup
Page-type segments, custom data extractions, and crawl-waste rules are generated automatically from your site's own data.
Every proposal is backed by executed queries you can inspect.
Ways to work with EdgeComet
- Dashboard - the full product at cloud.edgecomet.com.
- HTTP API - trigger cache refresh and invalidation from your CMS, deploy pipeline, or CI, and read your bot traffic and Search Console data as JSON or CSV.
- MCP server - connect Claude or any MCP client to your EdgeComet data: query it, research keywords, and stage Edge SEO changes from a chat.
Self-hosted engine
The rendering and caching core of EdgeComet is open source (Apache-2.0) and can be self-hosted. See the Self-hosted Engine documentation. The cloud platform documented here runs that same core as a managed service and adds the analytics, audit, and automation modules on top.