Mark Lavercombe
Mark Lavercombe is an individual/solo developer and creator of Agent Ready, an independent project that makes website AI-agent readability measurable and fixable. Agent Ready scans public URLs against the Vercel Agent Readability Spec, llmstxt.org, and agent-protocol standards, returning a score, per-check findings, and actionable remediation guidance.
At a Glance
- Website owners and operators
- Developers and engineering teams
- SEO/GEO and content teams concerned with AI visibility
- Teams implementing MCP, A2A, agents.json, UCP, x402, MPP, or NLWeb
- +1 more
AI Tools by Mark Lavercombe
(1)
Agent Ready
Website AI Agent Readiness Scanner
Discussions
No discussions yet
Be the first to start a discussion about Mark Lavercombe
Latest News
Agent Ready MCP repository updated its release workflow with a quote-safe mcp-publisher invocation fix.
Agent Ready CLI released at version 0.6.1.
Cursor plugin updated so scan_site works keyless on the anonymous free tier.
Agent Ready Gemini extension v0.1.1 added the accessibility sub-score to commands and documentation.
Products & Services
Web application that scans a public URL and scores it for AI-agent readability across the Vercel Agent Readability Spec, llmstxt.org, agent protocols, and a separate accessibility sub-score.
Versioned API for starting and polling scans, retrieving scan history, and querying Agent Ready methodology through the public ask/NLWeb endpoint.
MCP server exposing scan_site, get_scan, ask, and related validation capabilities for MCP clients; available as a hosted Streamable HTTP endpoint and an npm package.
Node command-line client installed from npm that provides scan, get, list, ask, MCP-server scan, and structured-schema validation commands.
Market Position
Agent Ready positions itself as an independent, protocol-focused alternative to platform-specific readiness scanners such as Vercel's is-agentic.com and Cloudflare's isitagentready.com, general SEO crawlers such as Lighthouse and Screaming Frog, and single-spec llms.txt validators. Its differentiator is a published, deterministic methodology spanning web readability, discovery artifacts, accessibility, and multiple agent-protocol surfaces, with API/CLI/MCP/SDK integrations.
Leadership
Founders
Mark Lavercombe
Solo developer who builds and maintains Agent Ready and publishes its surrounding open-source tooling under the GitHub account mlava. The official About page describes him as a single developer; the public GitHub profile identifies mlava as Mark Lavercombe.
Executive Team
Mark Lavercombe
Creator and maintainer
Solo developer behind Agent Ready; maintains the hosted service and its open-source CLI, MCP server, SDKs, extensions, and skills.
Founding Story
Agent Ready was started because the web was built for people to read while increasingly being consumed by AI agents and autonomous systems. Lavercombe's stated aim was to make that machine-facing readability gap measurable and fixable, rather than relying on manual audits or generic SEO tools.
Business Model
Revenue Model
Freemium subscription model: public website scanning is free, while Pro monetizes weekly monitoring, scan history, larger scan depth, REST API access, CLI access, email alerts, and authenticated MCP/CI capabilities. Agent-facing pay-per-scan micropayments are also supported through x402 and MPP.
Pricing Tiers
25 pages per scan, 10 signed-in scans per 30 days, keyless CLI and MCP scans within the anonymous quota, shareable result URLs, and badge embed.
250 pages per scan, 50 scans per month, 90 days of history, weekly monitoring, regression email alerts, REST API, CLI, authenticated MCP, GitHub Action/CI support, shareable result URLs, and badge embed.
$0.02 for a 25-page scan or $0.25 for a 250-page scan, settled on Base mainnet through x402 or MPP; this is an agent-facing payment path rather than a human subscription tier.
Target Markets
- Website owners and operators
- Developers and engineering teams
- SEO/GEO and content teams concerned with AI visibility
- Teams implementing MCP, A2A, agents.json, UCP, x402, MPP, or NLWeb
- Organizations using CI/CD and coding assistants such as Cursor, Claude, Gemini, Cline, Continue, or Windsurf
- Auditing a public website for AI-agent discoverability, parsing, and actionability
- Finding and fixing missing or invalid llms.txt, robots.txt, sitemap, structured-data, and agent-manifest artifacts
- Monitoring a website after deployments for regressions in agent readability
- Running repeatable readability checks from CI/CD, the command line, or an MCP-capable coding assistant
- Validating MCP, A2A, agents.json, agent-permissions.json, UCP, x402, MPP, and NLWeb implementations
- Improving how AI assistants cite, summarize, and interact with a site