Manifold MCP Review 2026: SEO, AI Search and Ads Data Inside Claude and ChatGPT

Some links on Digital Marketing Supermarket are affiliate links, and some listings and spotlights are sponsored. We may earn a commission or a fee from these at no extra cost to you. Sponsored placements are always labelled, and payment never decides a review score. We use AI tools to help with research and drafting, and an editor checks and edits everything before it’s published.

By Saj Ali · Last updated 3 Nov 2026

Manifold MCP is a single remote MCP server that gives AI assistants such as Claude, ChatGPT and Cursor live marketing data: Google search results, what AI engines say about a brand, Google Maps listings, ad libraries and social platforms. You connect it once, then ask questions in plain English. There’s a free tier with 250 credits a month and no card, and paid plans start at $20 a month (see Manifold’s pricing page, checked 3 November 2026). Our verdict: a well-designed, fairly priced tool for people who already work in an AI assistant, scored 4.2 out of 5: a solid, well-built tool from a very new, one-person business.

For the quick facts, see the Manifold MCP listing on Digital Marketing Supermarket. It overlaps with the AI visibility trackers we’ve reviewed, such as the Peec AI review and Profound review, but works very differently, and we explain how below.

Why marketing data is moving into the assistant

A growing number of marketers now do their thinking in Claude, ChatGPT or Cursor. The friction comes when the question needs real data: who ranks for this keyword, what ChatGPT says about us, which ads a competitor has run for months. Without a connection, you open another tool, export a spreadsheet and paste it back into the chat.

MCP, the Model Context Protocol, is the standard that fixes this. It lets an assistant call outside tools directly, so it can fetch the data itself and answer in the conversation. Manifold’s idea is to put many marketing data sources behind one MCP connection and one credit balance. If you’re new to working this way, our guide to using Claude skills for marketing and our piece on OpenClaw for digital marketers show what assistants can do once they have the right inputs, and you can browse other MCP-connected tools in our directory.

What is Manifold MCP?

Manifold describes itself as a hosted MCP server that gives AI agents live SEO, AI answer, social, ad and lead data. It was founded in September 2026 by Jesse Sibley, a growth engineer who says he has spent eight years in growth at London startups. Manifold says plainly that it’s a one-person company: the founder writes the code, runs the service and answers every support email (see Manifold’s about page). The operator is listed as Jesse Sibley, trading as Manifold, a sole trader in the United Kingdom.

Two clarifications help. Manifold is unrelated to Manifold Markets, the prediction market, and the site says so. And its data comes from DataForSEO and the public data of the platforms it covers, with Manifold’s own terms describing results as estimates and snapshots rather than guarantees.

Key features of Manifold MCP

One server, many assistants

The server sits at a single URL and works with Claude, Claude Code, ChatGPT, Codex, Cursor, Gemini CLI, Grok, OpenCode and any other MCP client (see Manifold’s docs). Most people sign in with OAuth 2.1 from the client, which opens a browser, creates a workspace on the free plan and needs no card. For scripts and clients that can’t sign in, you can use an API key, which is shown once and stored by Manifold only as a hash.

SEO data

The SEO tools cover Google organic data: keywords, rankings, results pages, domains, backlinks, competitor traffic estimates and a technical crawl. A typical prompt is “which keywords do we rank 5 to 15 for that could reach page one?”, the kind of question that normally means exporting from a paid SEO suite. If you use classic tools too, our list of the 15 best backlink analysis tools for 2026 and our 25 best AI SEO tools for 2026 show the alternatives.

Search Console, Bing and Google Analytics

Separate tool groups read your own Google Search Console, Bing Webmaster Tools and Google Analytics 4 data: clicks, rankings, sitemaps, indexing and crawl issues, plus sessions, conversions and revenue by channel. These cost no credits because they read your own accounts, though they are rate limited per workspace. They include sitemap and URL submission to Search Console and Bing.

AI answers

This is the part most relevant to AI search. The AI answer tools run a prompt across engines and report the brands each answer mentions, the sources it cites and the web searches the engine ran. According to Manifold’s docs, ChatGPT and Gemini answers are read live from the consumer apps, Claude and Perplexity answers come from their model APIs with web search on, and Google’s AI Overviews and AI Mode are covered too (see Manifold’s AI answers docs). A free site-readiness tool checks whether the engines can read your site.

Unlike a tracker, this is on demand. Manifold’s own about page makes the contrast: Profound, Peec and Otterly track the brands and prompts you set up in their dashboards, whereas with Manifold your agent can ask the engines about any brand and any prompt when the question comes up. That’s a vendor’s framing, but it’s an accurate description of the trade-off, and we cover it below. AI answers run as a background task and are never cached, because the point is what the engines say now.

Local, ads, social and leads

Beyond search, Manifold covers Google Maps businesses, Meta and Google ad libraries, Reddit, TikTok, Instagram, Facebook, YouTube and LinkedIn, app stores, Amazon and Google Shopping, and people and company lookups that can return a verified work email. For content and social teams, the ad-library and Reddit tools are the interesting ones: seeing which ads a competitor has run longest, and which threads ask for what you sell. Our roundups of the 25 best AI social media tools and best AI influencer marketing tools cover the workflow side.

Cost controls

This is the best-designed part of the product. Every result states what it cost. A dry_run option returns the estimate for free, and max_credits refuses any call whose estimate is higher. Manifold holds the maximum a call could cost, then charges only for what actually came back. A failed call, because a provider is down or the input is wrong, costs nothing, while a call that reaches a provider and finds nothing costs 1 credit (see Manifold’s credits documentation).

Caching and shared answers

Repeat questions are served from a shared cache of public data. The first cache hit in your workspace costs 25% of the live price, and after that the same answer is free until it expires. Cache lifetimes vary by data type, from an hour for Reddit and page lookups to a day for search results and 30 days for transcripts (see Manifold’s caching documentation). It’s a sensible design, but it means an answer may be hours or days old unless the tool is never cached, so check the data-as-of field before you report a number.

Limits and errors

Each workspace is limited to about 120 requests a minute and five calls in flight per provider family. Errors are typed, so an agent can react without guessing, and most cost nothing (see Manifold’s errors and limits page). You can also switch off tool groups you don’t need, which shortens the tool list and helps the assistant choose well.

How Manifold MCP works in practice

1. Add the server. Paste the URL into your MCP client, or follow the one-command setup for your client.
2. Sign in. Approve Manifold in your browser. That creates a free workspace with 250 credits.
3. Ask a real question. You don’t need to name tools. Try “what does Google show for ‘crm for startups’, and who ranks?”
4. Check the price. For anything expensive, ask for a dry run, and set a credit cap when you script it.
5. Verify and decide. Cross-check any number you plan to report, then choose whether a paid plan is worth it.

How much does Manifold MCP cost?

Manifold prices in credits, shared across every tool. There are no seats and no overage charges, plans are month to month and you can cancel at any time. The table shows what’s on Manifold’s pricing page on 3 November 2026.

What a credit buys

The pricing page gives these examples: the top 10 Google results for a keyword cost 1 credit, 100 keyword ideas cost 10, a domain overview costs 15, one prompt run on ChatGPT, Claude, Gemini and Perplexity costs 25, a 1,000-page technical crawl costs 50 and a work email costs 7. In the docs, AI answers cost 2 credits per prompt for ChatGPT, Gemini, AI Overviews and AI Mode, 6 for Perplexity and 15 for Claude, and the default of one prompt on five engines costs 27.

The small print on credits

Plan credits, including the free plan’s, reset each month and don’t roll over. Top-up credits never expire, and Manifold spends plan credits first. Top-ups cost $1 for 40 credits (2.5 cents each) and are available on paid plans. The FAQ says you can’t buy credits without a plan, so a free user who runs out waits for the monthly reset or upgrades. When credits run out, calls stop and nothing is charged. Auto top-up is available if you want it, but leave it off while you’re testing.

Is it good value?

For occasional research, very. The free tier covers 250 Google result lookups or ten four-engine AI answer checks each month, which is plenty to try the idea. The picture changes for tracking. At 25 credits per prompt across four engines, checking 50 prompts weekly would use about 5,000 credits a month, well above Pro’s smaller tiers. Dedicated trackers such as Peec AI are built for that job and price it flatly. Use Manifold for ad hoc questions and a tracker for the daily numbers.

Is the free tier enough to decide?

Yes, for the question that matters most: does having this data in your assistant change how you work? You can test it for free with no card, and a month’s credits are enough for a realistic trial. Run the same three tasks you’d normally do with a paid SEO suite, such as a keyword gap, a SERP check and an AI answer check, and compare the results with your usual tool. If the numbers differ, remember that all keyword volume and difficulty figures are estimates, whichever tool produces them.

Who Manifold MCP suits, and who it doesn’t

Good fit:

• SEO and growth marketers who already work in Claude, ChatGPT or Cursor
• Founders and small teams who want data on demand instead of a full suite subscription
• Freelancers and agencies prototyping research and reporting workflows with an assistant
• Outbound teams who want company and contact lookups beside their SEO data

Poor fit:

• Teams that need daily tracked dashboards and scheduled reports
• Enterprises that need an SLA, a security questionnaire or a vendor with a team behind it
• Anyone who wants a tool that publishes or acts for them, as Manifold looks things up

Pros and cons of Manifold MCP

Pros

• A free tier with 250 credits a month, month-to-month plans, no seats and no overage
• A price on every call, a free dry run and a hard cap per call
• Broad coverage in one connection: SEO, AI answers, Maps, ads, social and marketplaces
• Failed calls cost nothing, and repeat questions get cheaper through the cache
• Tools are read-only, apart from sitemap and URL submission to search engines

Cons

• A one-person company founded in September 2026, so it is early and unproven
• Data comes from third-party sources and is described as estimates and snapshots
• Heavy AI-answer checks use credits quickly
• It’s a data layer, not a tracker or an agent: you supply the questions and the judgement
• Plan credits don’t roll over, and the free plan can’t buy top-ups

Manifold’s tools read data rather than post, send messages or edit anything. The one kind of action it lists is sitemap and URL submission to Search Console and Bing. That’s small and doesn’t touch your content, but it’s worth knowing when you grant access.

Manifold MCP alternatives

Manifold sits between a classic SEO suite, an AI visibility tracker and a set of marketing skills. Prices are as shown on each vendor’s site when we last reviewed them; check before you buy.

Marketing Skills teaches an assistant how to do the work, while Manifold supplies the data to do it with, so the two combine well. For tracked AI visibility, read our Peec AI review and Profound review. For wider shortlists, see our guide to generative engine optimisation tools, our best AI marketing agents in 2026 roundup and our guide to AI agents for digital marketing. You can also browse the AI agents and AI search visibility tags.

Where Manifold fits in your marketing stack

Think of Manifold as a research layer, not a system of record. Use it for the questions you can’t predict: a competitor’s new ad campaign, what ChatGPT says about a brand you’ve just heard of, a one-off keyword gap before a pitch. Keep a tracker for the metrics you report every week, and your own Search Console and analytics for the numbers that matter most. For the wider picture of what assistants can do, see our best LLM for SEO in 2026 guide and our guides to answer engine optimisation and generative engine optimisation.

How we scored Manifold MCP

Every DMS review uses the same five criteria, each marked out of 5, and the overall score is a weighted average on the DMS 2026 scale, with features and evidence counting most. This is a desk review: we scored what the vendor publishes and what independent sources show, not a paid hands-on test.

Our verdict on Manifold MCP

Manifold MCP has a clear idea and executes the details well. One connection, broad data, a free tier, a price on every call and a hard spending cap are what an agent-friendly tool should look like, and the typed errors and caching show real thought. For a founder or SEO marketer who already lives in Claude or ChatGPT, it’s an easy way to add live data without another dashboard.

The caution is about the company, not the product. It’s a one-person business founded in September 2026, its data comes from third parties and is described as estimates, and there are no independent reviews yet. That’s why it scores 4.2 rather than higher. Start on the free tier, run three real tasks, and keep it as a research layer. If you need daily tracking, reporting or a vendor with a team, look at Peec AI or Omnia alongside it.

FAQs

What is Manifold MCP? Manifold MCP is a remote MCP server that gives AI assistants such as Claude, ChatGPT and Cursor SEO, AI answer, Google Maps, ad library and social data, paid per call in credits.

How much does Manifold MCP cost? On 3 November 2026, the free plan gave 250 credits a month with no card. Pro ran from $20 for 1,000 credits to $150 for 12,000, and Max from $250 for 21,000 credits to $2,000 for 200,000.

Is there a free plan? Yes. Free gives 250 credits every month with no card. Plan credits reset monthly and don’t roll over, and the free plan can’t buy top-ups.

Which AI assistants work with Manifold MCP? It works with Claude, Claude Code, ChatGPT, Codex, Cursor, Gemini CLI, Grok, OpenCode and other MCP clients, using OAuth 2.1 or an API key.

How do I control what Manifold spends? Ask for a dry run to get a free price estimate, and set max_credits to refuse any call that would cost more. Every result also states what it cost.

Does Manifold MCP change anything in my accounts? Its tools fetch data. The exception is sitemap and URL submission to Google Search Console and Bing, which are actions rather than lookups.

Can Manifold MCP track my AI visibility? It can check what ChatGPT, Claude, Gemini, Perplexity and Google’s AI answers say for any prompt, on demand. It doesn’t run a scheduled tracker, so use a dedicated tool for daily monitoring.

How much does an AI answer check cost? One prompt on ChatGPT, Claude, Gemini and Perplexity costs 25 credits. Per engine it is 2 credits for ChatGPT and Gemini, 6 for Perplexity and 15 for Claude.

Who makes Manifold MCP? Jesse Sibley, a UK sole trader based in London, founded it in September 2026 and runs it alone. It is unrelated to Manifold Markets.

This review is based on publicly available information from manifoldmcp.com and the sources linked above, checked between 25 October and 3 November 2026. We have not tested Manifold MCP hands-on. Digital Marketing Supermarket is an independent directory, and this review hasn’t been sponsored or paid for by Manifold. Vendor claims are reported as the vendor’s own.