Built on what won. Sharpened by what didn't.
Win Canadian grants using patterns from funded proposals.
Write your CCSIF proposal with AI grounded in three things at once: NSERC's official evaluation criteria, real funded applications, and the mistakes that got others rejected. Every section is scored as you write, then exported as an NSERC-formatted PDF for the Convergence Portal.
Three steps from idea to submission.
Start writing immediately
Pick your grant program (CCSIF template built in). AI drafts each section using patterns from winning grants. No profile setup required to get started.
Score & improve with AI
A 0–100 completeness score tells you how specific each section is. Competitiveness compares your writing to funded versus rejected proposals. Rewrite until every section is strong.
Formatted for Convergence Portal
Download properly formatted PDFs matching CCSIF Convergence standards: 11 pt Arial, letter size, correct headers. Ready to submit as-is.
How it was built
Grounded in three sources of truth, not generic AI.
Most AI writes grant-speak from thin air. CanGrant was built on real evidence: the funder's rulebook, the proposals that won, and the ones that lost. Every draft and every score obeys all three at once.
The rulebook
The funder's own instructions
NSERC's official CCSIF evaluation criteria and formatting rules, taken verbatim from their published guidance — exactly what reviewers score on, what disqualifies you, every page limit and EDI requirement. The AI follows this first, always.
What won
Real funded proposals
Real funded CCSIF applications, distilled into concrete patterns — numbered aims, named methodologies, living-wage citations, quantified partner contributions. The AI writes in the structure and language of grants that actually got funded.
What didn't
Lessons from rejected grants
Real rejected proposals and the reviewers' written feedback, distilled into the exact failure modes that sank them — vague EDI, deferred impact measures, unfocused scope. The AI actively steers around every one.
Winners show it the target. Losers show it the landmines. The funder's instructions are the ground truth. That is the whole idea behind Built on what won. Sharpened by what didn't.
Built on 4 funded CCSIF grants.
Every AI call, every score, and every suggestion is grounded in what actually won CCSIF funding. Here is a sample of the patterns the tool checks for: