GrantForge is the AI grant workspace for European R&I teams. It finds the calls that fit, evaluates your draft the way the panel will, and engineers the proposal with your consortium. You stay the author.
Where do you want to start? Or just ask me anything about EU grants.
Not autopilot. A workspace that runs every step a grant office actually struggles with, with you in control.
What I value most is how fast a researcher can test an idea against existing knowledge: the calls that fit, the projects already funded on the topic, the people who could join. With GrantForge, a viable proposition takes shape quickly, before the Go/NoGo.
The hardest part of the cycle is not the writing. It is not knowing what a reviewer will flag, until the rejection lands months later.
Import your Part B and get a panel-style evaluation in about 30 minutes. It surfaces the weaknesses that actually sink proposals, section by section, so you fix them while you still can. The score and the three most serious findings are free; the full report opens with a pass.
I sit in my university's AI group and I give AI workshops, so I am suspicious of AI in research. But what GrantForge is doing for ERC evaluation is a good use of AI. I'm on board.
Word, Excel and an uncontrolled ChatGPT do not catch when your Gantt contradicts your deliverables, and they do not hold a 15-partner consortium in sync.
A structured workspace where partners build async, the model stays connected end to end, and 23 live checks keep it coherent. It drafts section by section, you decide what stays. The result reads like your team, not like autopilot.
Both GPT and Claude are very useful, but they eventually reach a point where they stop improving the proposal meaningfully. They start polishing the same text unless the underlying project logic is strong. That is why GrantForge made a lot of sense to me. It addresses exactly the part where AI writing is not enough anymore, and where the real need is structured grant engineering.
A junior coordinator facing a blank 60-page Part B, and a support office that cannot be everywhere at once.
The Coach walks the team from call analysis to strategy to partner alignment, so no one starts from a blank page. Partners contribute async, the support office works from consolidated views with review built in, and one person can pilot the whole proposal. The team behind them does less of the low-value work.
It was my first European proposal, and a challenge to do it in 1 month instead of 5. It was quite demanding. In all honesty, without GrantForge I would never have done this proposal.
The point is not the features. It is what your week, and your win rate, look like once the busywork is gone.
Applications are up 60% since AI and scores are tighter. Fixing the weaknesses before you submit is how you keep your edge.
No more retyping references, chasing partner Word docs, or keeping a 60-page Part B in sync by hand.
Calls arrive in bursts. Consolidated views and a Coach that is always on let a support office cover far more proposals at once.
Confidentiality is the number one reason teams hesitate to put a grant in a tool. We built the stack around it, not as an afterthought.
A German cloud operator, not AWS, Azure or Google Cloud. No CLOUD Act exposure.
Your proposals are never used to train models. Mistral runs under Zero Data Retention, or bring your own key.
GDPR, a signable DPA, a full audit trail, passwordless sign-in.
Every step shares the same structured data, so nothing is retyped and nothing falls out of sync. That is the part no AI writer and no spreadsheet can give you.
Independent grant experts, specialist centres and research institutions across Europe work with GrantForge.
Armengaud Innovate
Amelia Tulacz
ECHO Science Communication Hub
University of Amsterdam