Due to the rise of agent-driven retrieval, the question of whether a website is machine-readable has shifted from an SEO concern to a survival concern. MachineRead is a free, anonymous audit tool that checks any public URL for the signals that AI agents, retrieval systems, and search crawlers actually use to discover, access, and understand a site. No account, no API key, no payment. You enter a URL and get a structured report.
This post covers what MachineRead is, why it exists, what it deliberately does not do, and how to run it. The reason for this shape is that most audit tools conflate measurement with recommendation, and the result is a report that tells you to fix everything without explaining which gaps actually constrain agent retrieval. MachineRead separates the measurement from the recommendation.
It is through the separation of measurement from recommendation that MachineRead avoids the trap of overclaiming. The audit runs 13 check groups covering 56 checked points across three pillars: off-site signals, AI access and scrapability, and SEO. Of the 13 check groups, 9 advanced rows are locked (not yet verified in Essentials, displayed as zero-scored coverage gaps rather than guessed at). The strict agent-readiness score exposes 8 default probes out of 21 full-scope probes, meaning the free audit already catches the majority of agent-native discovery signals without requiring paid crawlers or LLM verification. Four agent-consumable surfaces are checked directly: llms.txt, ai-catalog.json, api-catalog.json, and openapi.json.
MachineRead exists because the gap between "my site ranks on Google" and "an agent can retrieve my site's content" is widening. A site can rank well in traditional search and still fail agent retrieval entirely, due to missing structured data, blocked crawlers, or absent machine-readable API surfaces. The audit quantifies this gap. Percentages, integer counts, exact check names: the report is built on quantification rather than impression. The internal audit (MachineRead running against its own static export) passed at 5 pass, 5 partial, 3 fail-deferred, and 9 locked, which means the tool already detects its own coverage gaps honestly (the 3 fail-deferred results are transport artifacts of local HTTP testing, not product defects, and are documented in the public boundary notes).
What MachineRead deliberately does not do: it does not measure actual search ranking, traffic, backlinks, social traction, Core Web Vitals field data, model citation share, or conversion performance. It does not run paid crawlers, paid LLM integrations, or browser-agent task simulation. It does not make claims about what agents will do with your content once they retrieve it. The free Essentials audit measures the signals that constrain retrieval, not the downstream outcomes of that retrieval. This is a deliberate boundary, not a limitation to be fixed in a later version. As a result, the report is honest about what it can and cannot tell you.
The practitioner angle: MachineRead runs on a free-tier stack. The frontend is a static export hosted on Cloudflare Pages. The backend is a FastAPI app on the Render free tier. No paid APIs are called during a free audit. The cost ceiling is under $20 per month for the entire product, including mail. This is not a constraint that needs to be overcome; it is a design parameter. As a result, the audit is fast, bounded, and predictable. Meaning you can run it against your own site right now without signing up for anything. However, the free-tier posture also means the audit does not attempt deep crawling, browser rendering, or LLM-based content analysis, because those would require paid infrastructure that breaks the anonymous free-tier commitment.
So here is what I recommend. Enter your URL at machineread.ai, pick the preset that matches your site type, and run the audit. Read the report for the gaps it surfaces, not the score it assigns. The score is a convenience; the gap inventory is the product. If your site fails the agent-readiness probes, that is where to start. If it passes, the advanced rows show you what paid coverage would add. Either way, the report is free, anonymous, and takes under a minute to run.
The full blog rollout ships over the launch window, covering product ideology, methodology updates, and release notes as the surface evolves.
For the full methodology, see the methodology reference. For agent integration details, see the agent integration guide. Run the audit.