WebToMCP Review (2026): Pricing, Features & Honest Verdict

TLDR

WebToMCP crawls a website, indexes every page, and exposes it to AI clients like Claude, Cursor, and ChatGPT over the Model Context Protocol so they answer questions with cited sources. Setup is one step and the free tier is real. Best for: devs who want their docs searchable inside their AI assistant. Price: free plan, Hobby $9/mo. Rating: 7.3/10.

What is WebToMCP?

WebToMCP is a bridge between any public website and your AI assistant. You point it at a URL, it crawls and indexes all the pages, and it hands back an MCP endpoint. Connect that endpoint to Claude, ChatGPT, Cursor, or Gemini, and the assistant can suddenly answer detailed questions about that site, with citations back to the source pages. In practice it turns a documentation site, knowledge base, or company handbook into something your AI can read on demand.

We dug into it and the timing makes sense. The Model Context Protocol became the standard way to plug external data into AI clients, and most teams do not want to build and host their own MCP server just to make their docs queryable. WebToMCP removes that work. It handles the crawl, the indexing, the search infrastructure, and the protocol plumbing, then gives you a single endpoint. It is aimed at developers and small teams: people who want their docs, policies, or product knowledge available inside the AI tools they already use, without a custom backend.

The use cases are concrete. A developer can index a library’s documentation so Cursor stops hallucinating its API. A customer-success team can index company policies so support agents get accurate answers inside their assistant instead of digging through a wiki. A product manager can index the whole product site and ask questions across pages that would normally require ten browser tabs. In each case, the value is the same: the AI stops guessing about your content and starts answering from it, with a link back to the exact page.

What Are WebToMCP’s Key Features?

Automatic site indexing

WebToMCP crawls all public pages of a website and respects robots.txt. You give it a domain, and it builds a knowledge base from the content it finds. It also supports JavaScript-rendered content, which many simple crawlers miss, though that capability is metered: 5 JS pages a month on the free plan and 100 on Hobby. For static or server-rendered docs sites, the indexing is the part that just works.

Hybrid search on Cloudflare’s edge

Under the hood, WebToMCP uses a hybrid of vector search and BM25 keyword search, served from Cloudflare’s edge network. The hybrid approach matters: vector search catches semantic meaning while BM25 nails exact terms like error codes and API names, and running it on the edge keeps responses fast. The result is that the AI gets relevant chunks back quickly rather than guessing.

One-step AI client integration

Connecting WebToMCP to your AI client is a single step. It supports Claude, ChatGPT, Cursor, and Gemini directly, and there is a CLI for programmatic setup if you want to wire it into a script or onboarding flow. Because everything runs through the standard Model Context Protocol, you are not locked into a proprietary plugin format. Each endpoint is keyed to a unique URL for security.

Cited answers and auto-refresh

When the AI answers from a WebToMCP knowledge base, it returns cited sources, so you can click through and verify the claim rather than trusting a black box. This is the feature that makes the answers trustworthy enough to act on: in a developer context, a hallucinated API signature can waste an hour, so seeing the source page the answer came from matters. The index also refreshes automatically each week, with a manual re-crawl option when you ship a docs update and want it reflected immediately. Indexed content is encrypted at rest on Cloudflare infrastructure, and each knowledge base sits behind its own unique URL key, so endpoints are not guessable.

How Much Does WebToMCP Cost?

WebToMCP starts free. The free plan gives you about 2,000 questions a month, one knowledge base, 5 JavaScript pages a month, and weekly auto-refresh. That is not a trial, it is a usable tier for a solo developer or a single side project. No credit card is required to start indexing a public site.

The paid Hobby plan is $9/mo and raises the limits to roughly 5,000 questions a month, 5 knowledge bases, and 100 JavaScript pages a month. That covers an indie developer running several projects or a small team that wants a few separate docs sites indexed. The pricing is refreshingly simple: two tiers, clear limits, no enterprise sales maze.

The thing to size carefully is the JavaScript page limit. Most documentation sites are static or server-rendered, so the 5-per-month free and 100-per-month Hobby JS quotas never bite. But if your docs are a client-rendered single-page app, every page counts against that quota, and you can exhaust it fast. Check how your site renders before assuming the free plan will cover a full crawl.

Against alternatives, that is cheap. Mintlify’s AI features and Inkeep both sit in higher, often custom or team-priced ranges aimed at companies, while building your own crawler plus an MCP server costs you engineering time even if the tools are free. WebToMCP’s $9 ceiling for the documented use case is hard to beat unless your needs grow past two knowledge bases and light JS rendering.

Who is WebToMCP Best For?

Use WebToMCP if you want your public documentation searchable from inside Claude, Cursor, ChatGPT, or Gemini and you do not want to build an MCP server yourself. It is ideal for solo developers, indie hackers, and small teams: the free tier covers a single project, and $9/mo covers a handful. If your docs are static or server-rendered, you will barely notice the JS-page limits.

It is a particularly good fit for open-source maintainers and SaaS companies with public docs. Pointing Cursor or Claude at your indexed documentation cuts down on support questions and stops AI assistants from inventing API methods that do not exist. The cited-source feature means a contributor or customer can trust the answer enough to act on it.

Skip WebToMCP if your documentation is a heavy JavaScript single-page app, because the metered JS rendering will throttle you fast. Skip it too if you need private or gated content indexed, since it only crawls public pages, or if you want more than retrieval: WebToMCP answers questions, it does not write code, take actions, or manage content. For those jobs you need a broader platform.

Best WebToMCP Alternatives

Mintlify

Mintlify is a hosted documentation platform that layers an AI assistant and search on top of the docs it hosts. If you want your documentation site and its AI search to live in one place, Mintlify is the polished, integrated option. It is more expensive and assumes you host your docs with them, whereas WebToMCP works against any public site you already run.

Inkeep

Inkeep is a docs-focused AI search and MCP provider built for support and developer-experience teams. It goes deeper on analytics, answer quality tuning, and support integrations than WebToMCP. The trade-off is price and complexity: Inkeep is a team purchase, while WebToMCP is a $9 self-serve tool. Pick Inkeep if you are a company supporting users at scale.

Self-hosted crawler plus MCP server

The DIY route uses a crawler like Apify or Firecrawl to scrape a site, your own vector store, and a custom MCP server. This gives you full control over private content, refresh cadence, and search tuning, and the underlying tools can be cheap or free. The cost is your engineering time and ongoing maintenance, which is rarely worth it for a small project. Choose it only if you genuinely need the control over private content or search tuning that WebToMCP cannot give you.

Final Verdict: Is WebToMCP Worth It?

WebToMCP earns a 7.3. It does a narrow job cleanly: turn a website into an MCP knowledge base your AI assistant can search and cite. The free tier is genuinely useful, setup is one step, the cited answers build trust, and $9/mo is a fair price for the upgrade. For a developer who just wants their docs available inside Claude or Cursor, it is close to the fastest path there is, and the zero-cost starting point means there is no reason not to try it on a project today.

It does not score higher because it is deliberately limited. Metered JavaScript rendering, public-only indexing, and a retrieval-only feature set mean it will not stretch to cover heavy SPA docs, private knowledge bases, or anything beyond answering questions. Within its lane, though, WebToMCP is a smart, cheap, recommendable tool. If you outgrow it, Inkeep or a self-hosted MCP server are the next steps up.