Reviewer intelligence for grant writers

See the objections real evaluators raise on proposals like yours — before you submit.

Whetmark reviews your draft against the objections real EU evaluators wrote on proposals like yours — grounded in a private corpus of official reviewer reports — then marks the fixes into your Word document as tracked changes. On a private instance only you can reach; no public AI ever sees it.

Request a demo Watch the 21-second walkthrough

Private instance · only you can reach it No cloud AI, ever EU-hosted

● Sample pre-submission read

What real reviewers flag on sections like this

The specific concerns real EU evaluators raised on sections like yours — each with the passage and a proposed fix.

Implementation

concerns matched

Excellence

concerns matched

Impact

concerns matched

Matched by similarity against 81 real EU evaluations — the reports evaluators wrote on comparable proposals — to prompt your review, not to predict a score. Analysed on your own private instance.

The core of Whetmark

It has read the reviews — not just the rules.

Whetmark's edge isn't a bigger AI — it's what the AI has read. Its review is grounded in the official evaluation reports written by the EU's own expert evaluators: the exact, criterion-by-criterion criticisms — and the strengths — that decided real proposals, funded and rejected alike. No general-purpose tool has ever seen them — which is why it can show you the specific objections real evaluators actually raised on sections like yours. The corpus grows as we add more evaluations — never yours: your work is never added to it or used to train anything.

81real EU evaluations

602real reviewer criticisms

596real reviewer strengths

Current release: Reviewer corpus 2026.08f — filtered and de-duplicated from 931 criticisms and 681 strengths extracted from the reports. Every instance shows the release it is running.

The gap Whetmark closes

You write for the call. Reviewers judge you against every proposal they've scored before.

Most tools polish your grammar. Whetmark does something no general tool can: it surfaces the objections real EU evaluators most often raise on sections like yours, while you can still fix them — because we paired real grant proposals with the official reports their evaluators wrote, and taught the tool to spot the same problems in your draft.

Surface the objections evaluators actually raise

For every section, Whetmark surfaces the specific objections real EU evaluators raised on similar work. It draws on 602 criticisms from 81 real EU evaluations, maps each to the criterion a Horizon or national panel scores it under — Excellence, Impact or Implementation — quotes the passage that prompts the concern, and names the objection.

Fix it in your document

Accept the suggestions you want and Whetmark writes them straight into your proposal as Word tracked changes — not a list to retype. You stay in control of every edit.

See whether your revision answered each concern

Re-run the review on your next version and Whetmark compares it to the last, concern by concern: a local comparator flags each objection as improved, still open, or regressed — and it catches content you deleted, because more words don't count, only concrete substance does. You confirm every verdict; the tool defers the uncertain calls to you, and you submit once you're satisfied you've addressed what matters.

One proposal, a full panel

Not one AI opinion — a panel of specialist reviewers.

When you run a review, Whetmark doesn't make a single pass. It convenes a panel of expert agents — each one a mock evaluator for a different thing a real panel scores — and they read your whole proposal together, the way a Horizon panel splits the work between experts.

Three criterion reviewers

Excellence · Impact · Implementation. Each critiques its own criterion exactly as an evaluator would — it names the specific weakness, quotes the passage, and explains the objection, grounded in the real reviewer corpus and your own budget, partner and ethics documents. Then it proposes the fix.

The panel simulation

Reads the proposal as a whole and surfaces the cross-cutting objections real evaluators have most often raised on proposals like yours — the concerns that span sections, not just one paragraph at a time.

The checks

Template · Consistency · Clarity. Are all the required sections present? Do your claims, figures and terms line up across sixty pages? Is it written to be scored, not just read?

Intake, first

Before any of that, it flags the project information that's still missing — so you close the gaps before they cost you points.

They run as one comprehensive sweep — every agent reads the entire document, not just the opening pages — and hand back two things: the specific objections real evaluators most often raise on sections like yours, each grounded in the corpus, and a set of evaluator-grade suggestions you review, accept, and download as Word tracked changes.

See it in action

A real review — flagged, then rewritten.

An actual Whetmark review of a sample Horizon proposal. It read the whole draft against 602 real criticisms that EU evaluators raised on comparable proposals — across all three criteria, Excellence, Impact and Implementation — then returned eight specific, evaluator-grade fixes. Here is the workspace, and one of the eight.

A real review, end to end — upload → the objections real evaluators raise on sections like yours → the evaluator-grade suggestions → the tracked-change fix written into your Word document, then the writer-confirmed improvement gauge as you revise.

Your passage

“The risk register (Table 3.2) is reviewed at each monthly teleconference and updated by WP leaders.”

What the reviewer agent flagged

The proposal lacks a clear description of the process for escalating identified risks beyond monthly teleconferences — potentially hindering effective risk mitigation.

The tracked change Whetmark wrote

“…reviewed at each monthly teleconference and updated by WP leaders. In the event of high-priority risks, a dedicated risk-mitigation meeting will be convened with all relevant stakeholders to develop and implement a comprehensive action plan.”

One of eight from a single sweep. Book a walkthrough on your own proposal →

The non-negotiable

Your unpublished proposal never touches a public AI.

Whetmark runs on a private instance provisioned for you alone. The reviewer AI runs isolated on that instance — no OpenAI, no Gemini, no cloud model is ever called with your ideas, your consortium, or your budget. The confidentiality that a competitive proposal demands is built into the architecture — not a setting you have to trust.

✓ A dedicated, single-tenant EU instance — your data is never mixed with anyone else's.
✓ The reviewer AI runs on your instance; no third-party or cloud model is ever called.
✓ You are the only user of your instance. We connect only for support, with your permission — and every connection is logged.

Why this matters for grants

A grant proposal is your most sensitive unpublished work — novel science, a costed plan, a named consortium. The moment it touches a third-party AI it can be logged, cached, or trained on. Whetmark removes that risk entirely: there is no third party.

Built for Horizon Europe, ERC, MSCA, FWO & national calls.

A method, not just a tool

Six phases, from “should we bid?” to submission.

Whetmark is organised around the workflow in the companion playbook — so the tool and the method reinforce each other.

Qualify & set up

Decide to bid; plan the deadline.

Intake & safety

Bring documents in, locally.

Decode & frame

Turn the call into a map.

Design & draft

Impact pathway and the work.

Review & iterate

The mock-evaluator pass.

Finalize & submit

Tracked changes, final checks.

See what it costs Watch the walkthrough

The companion playbook · out now

The EU Grant Writer's AI Playbook

by Dr Pascal Kahlem · 18 recipes for Horizon Europe & ERC, grounded in real EU evaluation reports

The practical method behind Whetmark's phases — written for professional grant writers and research offices. Out now in paperback and on Kindle.

Kindle on Amazon → Paperback →

✓ Included free with every Whetmark licence

Method + machine

The book teaches the craft. Whetmark runs it.

Every recipe in the playbook has a home in the workspace — so a team learns one method and executes it in one place. New clients get the book as part of onboarding.

$Dr Pascal Kahlem, founder of Whetmark$

Who's behind Whetmark

Built by a grant writer — for grant writers.

I'm Dr Pascal Kahlem. I've spent years writing and reviewing EU research proposals, and I built Whetmark around the one source no general grant tool can draw on: the evaluators' own reports — the 602 real objections and 596 strengths EU panels wrote across 81 evaluations. I was tired of guessing what a panel would object to — so I taught a private AI to surface those same objections on a new draft, and to keep every word on an instance only you can reach. I also wrote the companion playbook it runs on. If you write proposals for a living, I built this for you.

Connect on LinkedIn

Pricing

One product. Everything included.

Your own private, EU-hosted Whetmark instance — we provision and manage it, no installation, no cloud AI ever. One simple licence — annual, or month-to-month with no commitment — invoiced by Scientific Network Management S.L.

30-day instance

€490/30 days

Your own instance for one submission cycle. Credited in full against a licence.

Prove it on your own real proposal, on your own side, before committing to a year.

✓ A private instance, provisioned for you, live for 30 days
✓ Everything the annual licence does — full sweep, objections, tracked-changes export
✓ Your proposal stays on your instance. We never receive it, so nothing to send and nothing to trust us with
✓ Enough time for one real submission: review, revise, re-check
✓ The €490 comes off the licence price if you continue within 60 days

One proposal, one full cycle. Credited in full if you continue.

Start a 30-day instance

Launch offer

Private instance

€2,400 €1,500/year

First year at the launch price for our first 10 founding clients; renews at €2,400. Or €250/month, billed monthly, cancel anytime.

Built for the grant writer — solo consultants and research-office officers alike.

✓ Your own private, EU-hosted instance
✓ Private reviewer AI — no cloud model ever sees your proposal
✓ Corpus-grounded reviewer objections — the specific concerns real EC evaluators raised on sections like yours, drawn from 602 real criticisms across 81 real EU evaluations
✓ Writer-confirmed improvement gauge — concern by concern, whether your revision actually addressed each objection (you confirm every call)
✓ Full multi-agent review sweep
✓ Tracked-changes .docx export
✓ The companion playbook
✓ Setup, updates & support included

Everything above, on your own instance, for a year.

Request a demo

Straight answers

What Whetmark will not do.

Every grant tool you have been pitched claims certainty. Here is where this one stops — stated before you buy, not after.

It will not predict your score.

It retrieves the objections real evaluators raised on comparable passages — a prompt to check, not a prediction. Panels are human and the competition is hidden. Distrust any tool that says otherwise.

It will not replace a professional reviewer.

The two are complementary. It takes the coverage work off that reviewer's plate — every section, every draft, against what evaluators actually wrote — so their hours go where judgement is genuinely needed.

It will not see your proposal.

Your draft stays on your own EU-hosted instance. We never receive it, so there is nothing to send us and nothing to trust us with. No public AI is ever called.

It will not lock you in.

Take one proposal through one full cycle for €490. If it does not earn its place in your process, you stop and owe nothing further. If it does, the €490 comes off the licence.

Try it on a real call

See your next proposal through a reviewer's eyes.

Book a walkthrough, or ask for a private instance for your team. We'll reply within one working day.