FullEnrich MCP Review 2026: Features, Setup and Limitations

Key takeaways:

• FullEnrich MCP connects Claude, ChatGPT and other clients to a waterfall that queries 20+ enrichment providers in sequence, over OAuth with no API key to manage.
• Search is free to explore. You get the first 10 results plus the total count at no cost, and credits only apply when you export, at 0.25 per contact or company.
• Enrichment costs the same through MCP as through the app or API, and the assistant is designed to confirm with you before spending anything.
• Credits are weighted by data type: a work email costs 1 credit, a personal email 3, and a mobile number 10.

Most enrichment tools charge you money to find out whether they can help you at all. FullEnrich does not, and that single pricing decision is the most interesting thing about its MCP server.

You can run a search, see how many people match and preview the first ten, then refine your filters as many times as you like without spending a credit. Payment starts when you export.

This review covers the tool surface, the setup, the credit mechanics and the limits. Outreach is our patch rather than enrichment, and we maintain a connector for it, so we will be clear about which half is which.

What FullEnrich MCP Is

FullEnrich MCP is the company’s own Model Context Protocol connector, using the vendor-neutral standard that lets assistants reach software they did not ship with. It puts FullEnrich’s contact search and enrichment in front of any compatible client.

The underlying product is a waterfall. Rather than querying one data vendor and accepting whatever comes back, FullEnrich runs a request through more than 20 providers in sequence until verified data is found, which the company reports as an 80%+ fill rate. That architecture is the reason to use it over a single-source tool, and the MCP server inherits it unchanged.

Authentication is OAuth. You are redirected to sign in with your FullEnrich account in the browser, you authorize access, and the session is stored by the client. No API keys and no manual configuration are required for the supported assistants.

That choice puts FullEnrich on the safer side of a line this category is still arguing over. Several outreach and data vendors still ask you to paste a long-lived key into a config file, and a few pass it in the URL. An OAuth session that the client refreshes on your behalf leaves nothing sitting in a text file for someone to find later.

The waterfall is also what separates this from the single-source email enrichment APIs that dominate the cheaper end of the market. Those return nothing when their one database has a gap, which reads to a rep as “this person has no email” rather than “this vendor could not find it.” Querying twenty providers in sequence turns most of those misses into hits, and it is the reason a waterfall costs more per record than a single lookup.

What You Can Actually Do

Three jobs, and they are designed to run in that order.

FullEnrich’s own example prompts show the intended shape: “Find me marketing directors at software companies in the US,” then “Enrich these 50 contacts with work emails and phone numbers,” then “Export these results as a CSV.”

Batch enrichment is the part that makes the FullEnrich MCP server more than a lookup toy. Asking for fifty contacts in one instruction is a normal request rather than an abuse of the interface, and it is the shape most prospecting work actually takes.

How Credits Work

This is the section worth reading twice, because the model is genuinely different from its competitors and the difference is in your favor.

Search is exploratory and free. Running a search returns the first ten results and the total count at no cost. FullEnrich’s stated reasoning is that you might need three or four attempts to get filters right, and none of that should cost anything. Credits only apply when you pull actual data out.

Export is where you pay, at 0.25 credits per contact or company, the same rate as the API. There is no MCP surcharge.

Enrichment is priced by what you ask for, which is the part people miscalculate. The three field types sit an order of magnitude apart: work addresses at the base rate, personal addresses at three times that, mobile numbers at ten. Requesting phone numbers across a large list therefore costs roughly ten times the same list in work emails, so the shape of your request matters more than its size.

At the published rate of around $0.055 per credit on a 1,000-credit monthly plan, that puts a work email near six cents and a mobile number near fifty-five. FullEnrich starts you with 50 free credits and no credit card, which is enough to test real queries against your own market before deciding anything.

One behavior is designed in and worth relying on: the assistant confirms with you before running enrichment, so nothing gets charged without your go-ahead.

Compare that against the norm before deciding it is unremarkable. Most credit-metered tools charge for the attempt, which quietly penalizes exploration and pushes teams to write over-broad filters so that one expensive query has to do.

Charging on export instead means the cost lands on data you have decided you want, and it changes how people actually use the tool. It is the most buyer-friendly model we have seen across the AI prospecting tools we have reviewed this year.

How to Set Up FullEnrich MCP

Setup is short because OAuth does most of the work for you.

Claude and ChatGPT

Both use a custom connector and a browser sign-in, so the sequence is the same in either client.

1. Open your assistant’s settings and go to the connectors section.
2. Add FullEnrich as a custom connector.
3. Complete the browser sign-in with your FullEnrich account and authorize access.

FullEnrich publishes separate step-by-step walkthroughs for Claude and ChatGPT, which are worth following rather than improvising, since connector menus move between client versions.

Claude Code and Other IDEs

Developer clients take the server through an MCP configuration block rather than a settings screen. The same OAuth flow then runs on first use.

Optional: The Skills Library

FullEnrich maintains an open-source library of skills that add higher-level guidance on top of the MCP tools, covering smarter confirmation flows, better search strategies and workflow templates.

The skills are genuinely optional and the server works without them. They matter most for repeatable multi-step prospecting, and FullEnrich also points to them as the fix if your assistant is skipping the confirmation step before enrichment.

If Something Goes Wrong

FullEnrich’s documented smoke test is the fastest diagnostic in this whole category: ask “Check my FullEnrich credits.” If that returns a number, the connection is fine and the problem is your prompt rather than your setup.

Two other documented quirks are worth knowing. Some clients do not auto-close the sign-in tab, so close it manually after the confirmation page appears. And if enrichment runs without asking first, being explicit (“show me the results before enriching”) restores the intended preview-confirm-enrich order.

Thin results usually mean over-tight filters rather than thin data. FullEnrich’s guidance is to broaden the criteria and use standard industry and job-title language, then narrow down, which is the opposite of how most people write a first query.

Non-standard titles are the common culprit, since a database indexes what companies actually publish rather than what a rep would call the role. The same discipline applies across B2B data APIs generally, where filter vocabulary decides hit rate more often than coverage does.

Where FullEnrich MCP Falls Short

The limits are mostly matters of scope, and each one follows directly from what the product is built to do.

Coverage is a claim you should test rather than accept, and FullEnrich makes that unusually easy to do. An 80%+ fill rate is a strong number and it is an average across everyone’s queries, not a promise about your segment. The free preview exists precisely so you can check it, and the 50 starting credits are enough to sample a real list.

Search quality depends on vocabulary more than most buyers expect, which the free preview makes cheap to discover but does not solve. A filter written in your internal language for a role will underperform one written the way companies title it, and no amount of credits fixes a mismatched query.

Phone numbers are expensive relative to email, at ten credits against one. That is honest pricing rather than a trick, since mobile numbers are genuinely harder to source, but a workflow that requests them by default will burn a plan far faster than the headline credit count suggests.

There is no outreach surface at all, and that is the boundary that matters most for anyone reading this. FullEnrich finds and verifies; it does not send, sequence or track replies.

Nor is there a CRM sync in the MCP flow. The documented output is an export, as CSV or JSON, which your assistant can parse and hand onward. That is flexible and it is also a manual seam: nothing writes the enriched record into your systems unless you build that step, so a workflow that runs weekly needs somewhere for the file to go.

Batch behavior deserves one caution. Asking to enrich a large list is a single instruction with a large bill attached, and the confirmation step is the only thing standing between a loose prompt and a significant spend. Naming the number, as in “enrich the first fifty,” is a better habit than “enrich all of these.”

Pricing predictability is the other soft limit. Because cost depends on which fields you request rather than how many records you touch, two months with identical record counts can produce very different invoices. Teams that want a flat monthly number tend to prefer a platform with a bundled database and accept narrower coverage in exchange for a bill they can forecast.

The compliance work stays with you, as it does with every enrichment vendor. FullEnrich is SOC 2 Type II compliant and publishes a trust center, which covers how it handles data rather than how you use it. The NIST Privacy Framework is the useful reference for that second half, being a tool to help organizations manage privacy risk through enterprise risk management rather than a checklist a supplier can pass on your behalf.

One technical footnote matters more than it looks. A “found” address is only useful if it is both syntactically valid and deliverable, and those are different things. RFC 5322 defines the Internet Message Format and the address syntax an address must satisfy to be well formed at all, which is the floor rather than the goal.

A waterfall that returns a well-formed address to a mailbox that no longer exists has still failed. That is why verification and enrichment stay separate steps in any serious stack, and why a fill rate is not the same measurement as a deliverability rate.

Pairing FullEnrich MCP With Reply MCP

FullEnrich answers who to contact and how to reach them. We answer what to say and when. The two servers meet at a single handoff.

The handoff mechanics are the part worth planning, and ours are specific enough to design around. Creating a contact needs a name plus either an email address or a LinkedIn URL, which maps cleanly onto what a FullEnrich export contains. Enrollment then takes up to 100 existing contacts into a sequence in a single call, so a hundred-row export becomes one instruction rather than a hundred.

That batch ceiling is the number to build against. A 500-contact list is five enrollment calls, not one, and an assistant that does not know this will either stall or silently truncate. Telling it the limit up front produces better behavior than letting it discover the boundary.

Our own gap sits exactly where FullEnrich’s strength is, and we would rather name it than let it surprise you. We expose no enrichment, people-search or email-finding tools at all, and contact lookup on our side reads the workspace you have already built rather than any outside database. Sourcing is FullEnrich’s job here and not ours.

Teams comparing the sourcing layer more broadly will find our roundup of contact enrichment APIs covers the field, including several vendors whose waterfall approach is similar and whose pricing is not.

Who FullEnrich MCP Is For

FullEnrich MCP suits teams whose bottleneck is contact data rather than sending, and it suits careful buyers particularly well. The free search preview and the 50 starting credits mean you can evaluate coverage on your own market before committing. Others offer a way in too, whether a free tier or trial credits, but paying only at the point of export is a narrower and more useful promise than a general allowance.

High-volume phone prospecting is the less comfortable fit, simply because mobile numbers are ten credits each and that arithmetic scales quickly. A thousand mobile numbers is ten thousand credits, which is an order of magnitude above what the same list costs in work emails. It is also the wrong tool if you expected an all-in-one platform, since everything after the export happens elsewhere.

Teams weighing where enrichment should sit in the stack will find our guide to lead enrichment APIs useful for the build-versus-connect decision.

The question that usually settles it is whether enrichment is a burst or a habit. Campaign-shaped demand suits per-credit pricing, because you pay for the months you actually prospect. Steady, low-level enrichment often costs less inside a subscription that bundles a database, and the comparison is worth running at your real monthly volume rather than at a list price.

List enriched, campaign still to build? Start a free trial and wire the two together.