DIRECT ANSWER / 15
What are the limits of this grant cross-checker?
It performs arithmetic and exact textual cross-reference on material you supply. It cannot interpret funder instructions, judge eligibility, allowability or reasonableness, read rich document formats, write or score proposal text, or approve anything.
Why this review exists
Every finding the tool produces is a statement about two documents you gave it, and being explicit about that is what makes the findings usable.
What it will not interpret
There is no funder-rule engine. Eligibility, allowability, cost principles, matching requirements, indirect-rate agreements and submission rules are all outside the product, and it will not tell you whether a cost qualifies.
It does not score, rank or assess a proposal in any way, and it refuses semantic judgement entirely. A mapping is a numeric equality that a person confirmed, not an opinion about whether the narrative justifies the cost.
It will not read PDF, DOCX or spreadsheet binaries. Strict delimited text and plain prose are the accepted inputs, because a tolerant parser for rich formats is where a silent misreading would come from.
What it will not assert
A freeze is not a signature, an approval or a certification, and a checksum establishes byte integrity and nothing about authorship or authority.
The tool cannot confirm that the budget file you imported is the one that will be submitted, or that the narrative you pasted is current. Recording the version labels is the whole of its provenance claim.
What remains is genuinely useful: exact, reproducible, traceable arithmetic and a per-row record of what a reviewer confirmed, refused and asked about, in a form that can be handed to somebody else.
Practical sequence
- Read this page before a real review
- Apply the controlling funder instructions yourself
- Have a person confirm every mapping
- Treat a quiet review as a starting point, not a result
What remains outside the app
Grantweft performs arithmetic and textual cross-reference on user-supplied material. It does not write proposals or determine eligibility, allowability, accounting treatment, compliance, completeness, accuracy, submission readiness or funding outcome.
The private desk keeps exact source labels, stable budget row identities, paragraph-based claims, confirmed mappings, visible arithmetic, findings, human decisions, review freezes and passive handoff files connected without an account or remote application database.