HomeAI Detector › For Grant Writers

AI Detector for grant writers, built for proposals, funder reports, and impact statements.

Pre-scan every narrative section before the grant office reads it, and before the package leaves for NIH, NSF, Wellcome, MacArthur, Ford, Open Philanthropy, or the Gates Foundation. Sentence-level highlights show exactly which lines react AI on methodology, theory of change, need statement, and sustainability prose. Calibrated for long-form professional grant narrative. Team workspaces on Business. GDPR aware, no training on submitted drafts. Free to try. No card.

Start free, no card See pricing
Pro at $19.99/mo, $14.99 yearly GDPR aware 90-day proposal history
Who it is for

Built for grant proposals, funder reports, and impact statements.

Professional grant writers ship 20 to 50 applications a year across federal agencies, private foundations, community funders, and corporate giving programmes. The pre-submission scan that fits a federal NIH narrative rarely fits a community foundation letter of inquiry without adjustment, and the same workflow runs again at the funder report stage twelve months later.

The grant writing stack runs from the concept memo to the close-out report. Pre-scanning fits every stage because reviewer panels at major funders now read with AI in mind, and several agencies have added human authorship attestation language to submission policy as of 2025.

Federal proposals (NIH, NSF, NIH-NIAID, DOE)

Specific aims, significance, innovation, approach, and broader impacts each carry distinct AI risk profiles. Approach and methodology flag heavier because the structure overlaps with templated paraphrase. Significance and broader impacts attract stock phrasing under deadline pressure. Free tier covers a single narrative subsection up to 5,000 characters. Pro at $19.99 a month, or $14.99 a month on yearly, unlocks 10,000 character pastes and unlimited scans for the iteration weeks.

Foundation proposals (Wellcome, MacArthur, Ford, Open Phil, Gates)

Private foundation narratives reward organisational voice. A programme officer reading 40 letters of inquiry in a week recognises story-led prose because it feels written by a person who knows the organisation. Templated need statements and outcome tables erode that voice. The pre-submission scan catches the patterns before they reach the grants committee.

Funder reports and impact statements

The same workflow runs at the report stage. Annual reports, mid-grant updates, and close-out impact statements all carry the same AI risk and the same calibration targets as the original proposal. The 90-day Pro history keeps every report scan retrievable across the grant cycle, useful when a programme officer references an interim report in a renewal conversation.

Major funder context

NIH, NSF, Wellcome, MacArthur, and Ford now flag AI prose.

Federal agencies and major foundations updated submission policy through 2025 to require human authorship attestation or AI disclosure on the cover form. Reviewer panels read narratives with AI in mind regardless of policy text. Pre-scanning is the defensible posture in 2026.

NIH and NIH-NIAID

NIH updated policy in 2025 to require human authorship attestation on submitted research narratives, with NIH-NIAID adding explicit language flagging AI-generated content as a research integrity concern. Specific aims and approach sections attract the closest reviewer review. AI-templated approach paragraphs dilute the experimental hypothesis and reviewer panels notice within seconds.

NSF

NSF emphasises intellectual merit and broader impacts as dual review criteria. Broader impacts in particular attracts AI-templated phrasing because the section invites general claims about educational outcomes, public engagement, and underrepresented populations. Reviewers know the standard tropes and flag them quickly. The pre-submission audit catches stacked abstractions before they read as boilerplate.

Wellcome Trust and MacArthur

Wellcome Trust requires AI disclosure on the cover form for major research programmes. MacArthur asks applicants to attest to authorship on the application portal. Both reviewer panels include AI-shaped prose in their evaluation notes regardless of whether the formal policy text mentions it. The defensible posture is to pre-scan, rewrite reds, then disclose honestly.

Ford, Open Philanthropy, Gates Foundation

Major US foundations including Ford, Open Philanthropy, and the Gates Foundation added AI disclosure questions to their application portals during 2024 and 2025. Community foundations and smaller family foundations are following at varying speeds. Treat every funder as if AI disclosure matters and let the scan history be the evidence trail that supports the disclosure.

Federal versus foundation attestation

Federal agencies tend to be stricter, with explicit attestation language and integrity consequences for misrepresentation. Foundation policies are usually softer but the reviewer culture is harder to predict. The pre-submission scan is the same either way, but the disclosure posture differs: full attestation language at NIH and NSF, honest disclosure of AI tools used for outlining or synthesis at most foundations.

Plans & pricing

Pricing for solo writers and development teams.

Pro at $19.99 a month standard, $14.99 a month on yearly billing, fits solo grant writers and consultants. Business at $39.99 a month, or $29.99 a month on yearly, fits nonprofit development offices and grant-writing firms with 5 seats. Full details on the pricing page.

Free
$0/forever

 

Sample a single narrative section or letter of inquiry.
  • 3 scans / day
  • 5,000 chars per scan
  • Sentence-level highlights
  • 2 lifetime AI rewriter uses
Start free
Starter
$7.49/month

Billed $89.88/year — Save $30

For an in-house staff writer running one application a month.
  • 20 scans / day
  • 20,000 AI rewriter words/mo
  • Chrome extension
  • Email support
Get Starter
Business
$29.99/month

Billed $359.88/year — Save $120

For development teams, research administration, and grant-writing firms.
  • 100,000 AI rewriter words/mo
  • 5 team seats, shared history
  • Audit log, REST API
  • White-label PDFs
Get Business

Most consulting practices and development offices see Pro or Business pay for itself on the first proposal that would have been rejected on a reviewer AI flag. View full pricing →

Pre-submission workflow

Scan before institutional review board, grant office, and funder portal.

Most grant packages clear three checks before they reach the funder: the writer's own draft pass, the development office or IRB compliance review, and the institutional grant office sign-off. The pre-submission scan fits before the first internal review, so the version your colleagues read is already AI-clean.

Step 1: draft normally

Write in your usual editor: Word, Docs, or the funder portal. Using ChatGPT for an outline pass, a literature synthesis, or to break writer's block on a difficult section is the realistic 2026 default. Compose the prose itself in your own voice or the organisation's voice from your notes, prior funded proposals, and conversations with programme staff.

Step 2: paste and scan section by section

Open app.textsight.ai, paste each narrative section, and scan. Free tier handles 5,000 characters in one paste. Pro handles 10,000, which fits a typical narrative subsection. For a 10-page narrative, split by section: need statement, significance, methodology, evaluation, sustainability, broader impacts. The scan returns in well under a minute with an Authenticity Score and a sentence-by-sentence colour map.

Step 3: edit by section type

Need statement, impact, significance: above 75, submit; below 75, rewrite reds in organisational voice. Methodology, theory of change, lit review: above 55, the sentence map is the signal, not the score. Evaluation and sustainability: above 70, rewrite reds rather than humanising stock phrasings.

Step 4: hand the package to internal review

Send the consolidated package to your development director, IRB, or institutional grant office for compliance review. Pro history keeps every section scan for 90 days. PDF export gives you a contemporaneous record. A typical narrative subsection round-trips in about fifteen minutes; a full 10-page narrative in about ninety.

Common false-positive patterns

Methodology, lit review, and theory of change trigger more flags.

Three sections of any grant narrative carry the highest false-positive risk because the structure is templated by design. The headline score reads lower here than elsewhere; the sentence map is the diagnostic that matters. Calibrate your threshold by section, not across the whole narrative.

Methodology: aim 55 to 70

Methods is procedural by design. Sample descriptions, instrument descriptions, statistical reporting, and analytical procedures carry precise claims with limited room for voice. Expect lower scores than other sections and rely on the sentence map. Scattered yellows across a structured methods section are usually domain register, not residue. Red clusters in one paragraph are worth a careful rewrite.

Literature review: aim 55 to 70

Citation-dense scholarly paraphrase reads structurally close to AI paraphrase. Expect lower scores than the rest of the narrative and rely on the sentence-level highlights. Aim for the 55 to 70 band on lit review and use the sentence map for the rest. Restructure rather than humanise on red sentences in lit review prose.

Theory of change and logic model: aim 60 to 75

Outputs feeding into outcomes feeding into impact is templated by definition. The risk is identical clause structure across three paragraphs, which reads as generated even when each link in the chain is real. Vary the construction, embed concrete actor language, and let the sentence map confirm the rewrite cleared the threshold.

Need statement and impact: aim 75 plus

Need statements stacking three or four consecutive sentences each opening with a statistic and a clause about underserved populations read as generated paragraphs rather than authored narrative. Impact statements with tripled adjectives (comprehensive, evidence-based, community-driven) are an AI tic reviewers spot on a scroll-through. Aim 75 plus on both and rewrite reds rather than humanising.

Multi-author drafts

Co-PI inputs, programme staff, evaluator, and finance for budget narrative.

Most grant narratives ship from three to seven contributors. Each contributor enters AI assistance at a different point and the combined draft can read uneven, which itself reads as AI. Business tier with shared seats lets the development team see the same scan and agree the threshold as a team.

Lead PI, co-PIs, and programme staff

Lead PI typically owns specific aims, significance, and approach. Co-PIs contribute methodology and analytical sections. Programme staff add evaluation and outcome detail. Each voice is different and each enters AI assistance differently. The combined draft can read like five authors stitched together, which reviewer panels read as inconsistent and flag.

Evaluator and finance for budget narrative

The evaluator drafts evaluation design and outcome measurement language. Finance adds the budget narrative tying line items to programme activities. Both sections carry domain vocabulary that overlaps with templated AI phrasing, and both contributors are usually less practised at sentence-level voice than the lead PI. Pre-scan each subsection before consolidation.

Shared scan threshold on Business tier

Business tier with 5 seats and shared history lets the development director, grants manager, lead PI, and programme lead see the same scan and flagged sentences. The Authenticity Score becomes the team threshold rather than a per-author judgment. The grants manager runs the consolidated scan once the narrative is assembled, and the team rewrites red sentences against the shared bar.

Subcontractors and contracted writers

Grant consulting firms and large development offices routinely contract individual narrative sections to subcontractors. Business tier audit log records who scanned what and when, useful when a subcontractor delivers a section that flags higher than the rest of the narrative. The contemporaneous scan record gives the grants manager something specific to discuss in the revision conversation.

What you see in a scan

Sentence highlights, paragraph cards, perplexity, and burstiness.

A single percentage is not a fix path on a 10-page narrative under deadline. The TextSight result panel shows which sentences reacted and why, with paragraph-level rollups for long sections, so you can edit the specific lines instead of rewriting the whole submission.

Sentence-level highlights

Every sentence is colour-coded by its own AI-likeness score. Red sentences clustered in one paragraph are a stronger signal than scattered yellows. Scattered yellows in otherwise structured methods or theory-of-change prose often just mean the writing is formally taught. You read the pattern, not just the headline number.

Paragraph cards for long narrative sections

For a 9,000-character methodology section or a 7,000-character evaluation plan, paragraph-level rollups identify which sub-paragraphs drift AI-shaped and which stay clean. Useful when you have a long section and need to know which two paragraphs to revisit rather than rereading the whole section under deadline.

Perplexity, read-only on Pro

Perplexity is how predictable your word choices are to a language model. Low perplexity reads AI-like. The score is shown per sentence on Pro, which is the diagnostic context you need to decide whether a flag is real AI residue or just an unusually well-rehearsed sustainability paragraph.

Burstiness, read-only on Pro

Burstiness is how much your sentence length and structure vary across the section. ChatGPT defaults to uniform medium-length sentences. Real organisational writing has bursty rhythm: one short sentence, one long, one fragment, one parenthetical. Low burstiness across an entire narrative section is the classic AI fingerprint and the one reviewer panels learn to spot first.

90-day history on Pro

Every scan is retrievable for 90 days. For a grant writer iterating across a 6-week proposal cycle, that means every clean section scan and every revision is on record. PDF export lets you save longer-term archives proposal by proposal, beyond the 90-day window. Business tier offers indefinite retention for consulting firms keeping records by client.

Your grant content stays yours

Privacy first: no training on grant content, GDPR aware.

Grant narratives are unpublished organisational work. They are also covered by GDPR in the EU and UK and by local equivalents elsewhere. TextSight is designed to honour those rules out of the box, not as a paid setting you have to find. A standard DPA is available on Business and Enterprise tiers for institutional procurement.

No training on submitted text

Narrative drafts, funder reports, impact statements, and budget narrative submitted for scanning are never used to train the classifier or any other model. This is a contract clause, not a configuration toggle. It applies on the free tier the same way it applies on Pro and Business.

No account required on free

The free tier needs no email, no account, no identity. For grant writers working on confidential applications or pre-disclosure funder strategies, this matters. You can scan a draft section without TextSight ever knowing who you are or which organisation you write for.

Programme officers and reviewer panels do not see your scans

Scan history is private to your account. We do not share scan data with funders, programme officers, reviewer panels, institutional grant offices, IRBs, Turnitin, iThenticate, or any third party. Your scans are not part of any external record.

Deletion on request, DPA on Business

Any scan can be deleted from your history. On Pro you can delete individual records. Data retention is bound to your settings, and a standard DPA is available on Business and Enterprise tiers for nonprofit development offices and research administration teams that need one for institutional procurement.

FAQ

Grant writers frequently ask.

Do federal funders like NIH and NSF run AI detection on proposals?
NIH and NSF have both updated 2025 policy to require human authorship attestation on submitted narratives, and NIH-NIAID added explicit language flagging AI-generated content as a research integrity concern. Neither agency confirms automated AI detection on every submission, but reviewer panels increasingly note AI-shaped prose in their critiques. The defensive posture is to pre-scan every narrative section and rewrite flagged sentences before the package goes to the grant office for institutional sign-off.
How does TextSight handle methodology and theory-of-change sections?
Methodology, literature review, and theory of change sections are templated by design and flag heavier than the rest of the narrative. The sentence-level highlights are the diagnostic that matters here, not the headline score. Scattered yellows across a structured methods section are usually domain register, not AI residue. Clusters of red sentences in one paragraph are worth rewriting in your organisation's voice. Aim for an Authenticity Score in the 55 to 70 band on methods and theory of change, and 75 plus on need statement, impact, and significance.
How does TextSight fit a multi-author grant draft with co-PIs and subcontractors?
Most grant narratives ship from three to seven contributors: lead PI, co-PIs, programme staff, evaluator, finance for budget narrative, and sometimes a contracted writer. Each contributor enters AI assistance at a different point and the combined draft can read uneven, which itself reads as AI. Business tier with 5 seats and shared scan history lets the team see the same Authenticity Score and flagged sentences, so the threshold is a team agreement rather than a per-author judgment. The grants manager or development director runs the consolidated scan before institutional review.
What is the pre-submission workflow before grant office review?
Draft each narrative section normally. Paste into TextSight and scan section by section: need statement, significance, methodology, evaluation, sustainability, broader impacts. Pro caps each scan at 10,000 characters, about 1,600 words, which fits a typical narrative subsection. Rewrite red sentences in organisational voice and rescan until each section clears the threshold. Then hand the consolidated package to the grant office or IRB for compliance review. The TextSight scan history gives the development office a contemporaneous audit trail in case a programme officer asks downstream.
Which tier fits a solo grant writer versus a development team?
Pro at $19.99 a month, or $14.99 a month on yearly billing, fits the solo grant writer or consultant pushing 20 to 50 applications a year, with 10,000 character pastes, unlimited scans, 90-day history, and PDF export for client deliverables. Business at $39.99 a month, or $29.99 a month on yearly, fits nonprofit development offices, university research administration, and grant-writing firms with 5 seats, shared history, audit log, REST API, and white-label PDFs. One avoided rejection on a six-figure ask covers either tier for a long time.
Is federal grant attestation different from foundation grant policy?
Federal agencies including NIH, NSF, NIH-NIAID, DOE, and DARPA tend to be stricter, with explicit human authorship attestation clauses in submission policy as of 2025. Major foundations including Wellcome Trust, MacArthur, Ford, Open Philanthropy, and the Gates Foundation now ask AI disclosure on the cover form, and smaller community foundations are following. Smaller family foundations may be silent in policy but the reviewers still read with AI in mind. Treat every funder as if AI disclosure matters, and let the pre-submission scan be the evidence trail that supports an honest disclosure or a no-AI claim.
Does TextSight train on grant content or share drafts with anyone?
No. Narrative drafts submitted for scanning are never used to train the classifier or any other model. This is a contract clause, not a configuration toggle. We do not share scan data with funders, programme officers, reviewer panels, institutional grant offices, IRBs, or any third party. Data retention is bound to your history settings, deletion on request is supported, and our privacy practices are GDPR aware in the EU and UK and aligned with local equivalents elsewhere. Business and Enterprise tiers can sign a standard DPA for development offices that need one for institutional procurement.
What about false positives on impact statements and need statements?
Impact statements, need statements, and sustainability sections carry stock phrasing patterns that overlap with AI-templated language even when written from scratch. The TextSight classifier is calibrated for long-form professional prose including grant narrative, so domain vocabulary like logic models, theory of change, scalability, and sustainability does not penalise the score on its own. The sentence-level highlights surface the specific construction patterns that flag, so you can rewrite the lines rather than chase a single percentage.
Related

More for grant writers and development teams.

Pre-scan every narrative section. Submit clean. Disclose honestly.

Free to try. No card. Pro at $19.99/mo, or $14.99/mo on yearly. Business with 5 seats for development teams.

Start free, no card See pricing
GDPR aware · No training on grant content · Sentence-level highlights · Team workspaces on Business