NoAIReceipt

Who it's for

Indie SaaS and desktop-app vendors marketing to AI-skeptical customers; today they cope with README 'no AI' statements, manual dependency lists, and security-questionnaire spreadsheets.

The problem

Payment: lost or delayed sales when buyers doubt no-AI claims; time: hours answering questionnaires and hunting dependencies; legal: false-advertising risk if a hidden SDK calls an LLM API.

How to build it

CLI plus GitHub Action plus hosted verification page; inputs are a public repo, lockfiles, container image, or release archive; outputs JSON manifest, signed PDF attestation, and embeddable badge.

How it makes money

Vendors pay $29-$99/month for continuous per-release attestation and public verification because a free grep cannot keep detector rules current, triage false positives, or produce signed commit-level evidence buyers can check.

Why it doesn't exist yet

Incumbents build vulnerability/license scanners or enterprise AI governance suites and avoid negative no-AI claims because rules are fuzzy; an indie can fill the gap with a curated public ruleset of known AI SDKs, model files, and API hosts turned into a simple receipt.

First users

First 10 come from 'no AI' app lists, LibreOffice/AI-free discussions, and indie SaaS founders already putting no-AI claims in their README; offer a free scan that replaces the claim with a public receipt.

Build size

1 person x 8 weeks; includes npm/pip/cargo/go dependency scanning, file heuristics for model weights/prompt packs, known AI API hostname search, GitHub Action, signed badge/report page; excludes runtime egress monitoring, binary decompilation, and legal certification.

Biggest risk

GitHub, npm, or a major SCA vendor ships native AI-dependency labeling, or the no-AI marketing wave fades before enough vendors pay; false negatives would also undermine the receipt.

Conditions for a hit (all 3 required)

• Given a public repo URL or release archive, returns a JSON manifest of detected AI-related dependencies, model-weight-like files, and known AI API hostnames with file paths and rule IDs in under 30 minutes.
• For a passing commit/release, publishes a verification page with commit SHA, image digest if used, scanner version, ruleset version, timestamp, downloadable signed PDF, and an embeddable SVG badge.
• A GitHub Action named noai-receipt posts a pass/fail status check and comments the top five detections when a PR adds an AI-related dependency, file, or hardcoded AI endpoint.

How it's judged (in 6 months)

A matching public service/tool reaches Product Hunt daily top 5 or a matching GitHub repo reaches 1,000 stars; alternatively, GitHub/code search shows at least 50 public repos/sites embedding the noai-receipt badge.(judgment date 2027-03-12)

AI self-confidence 38/100 — self-reported likelihood of meeting the criterion, not a business success rate