# Preuve AI - Full Context

> Market research and competitive intelligence for founders, and the white-label engine agencies, accelerators and consultants use to pre-qualify inbound startups. Real signals, not a confident AI score with nothing behind it. A free scan in about 60 seconds, a full paid report in about ten minutes. Preuve AI uses 10 AI agents to scan 60+ live data sources in parallel. It returns a verdict that has to be earned (median 54/100, only 17.5% reach 70), mapped competitors, demand signals, blind spots, and next steps for founders and teams. Every paid report runs on Claude Opus 5.5, Anthropic's most capable model, with the verdict cross-checked across multiple runs.

Last updated: 2026-10-22
Canonical domain: https://preuve.ai
Canonical facts: [https://preuve.ai/facts](https://preuve.ai/facts) is the dated reference for what Preuve AI is, what it costs, how long it takes and what a report contains ([Markdown](https://preuve.ai/facts.md)).
Compact version: [https://preuve.ai/llms.txt](https://preuve.ai/llms.txt)

## Entity

- Brand: Preuve AI (pronounced "proov", rhymes with "groove"; French for "proof/evidence")
- Category: SaaS / Startup validation / Market intelligence
- Founded: 2026
- Founder: Vincent Forat ([@VincentBuilds on X](https://x.com/VincentBuilds); [LinkedIn](https://www.linkedin.com/in/vincent-forat/))
- [Contact](mailto:hello@preuve.ai): hello@preuve.ai
- [Trustpilot](https://www.trustpilot.com/review/preuve.ai): customer reviews
- [Product Hunt](https://www.producthunt.com/products/preuve-ai): launch profile
- [Wikidata (company)](https://www.wikidata.org/wiki/Q138665055): structured entity record
- [Wikidata (founder)](https://www.wikidata.org/wiki/Q140248049): structured entity record
- [Crunchbase](https://www.crunchbase.com/organization/preuve-ai): company profile
- [LinkedIn](https://www.linkedin.com/company/preuve-ai): company page
- [Reddit](https://www.reddit.com/r/ValidateThisIdea/): official community subreddit
- Previous brand: TestYourIdea / TestYourIdeaHQ, now redirected to preuve.ai

## Citation-ready facts

- 9,000+ startup ideas analyzed since launch.
- The brand name "Preuve" is French for "proof/evidence" and is pronounced "proov" (rhymes with "groove"). Previously branded TestYourIdea.
- 10 AI agents scan 60+ live data sources per analysis.
- The Preuve AI Score is a 0-to-100 viability rating built on six factors, each scored against cited live evidence. The median idea scores 54. A score above 70 indicates launch-ready signals. Below 40 means fundamental validation work remains. The score is computed by 10 AI agents cross-validating claims against 60+ live sources including competitors, demand signals, market size, community sentiment, hiring trends, and patent activity. See [How to Validate a Business Idea with AI (Without Trusting a Guess)](https://preuve.ai/blog/how-ai-validates-startup-ideas).
- Free scan returns results in about 60 seconds.
- In the July 2026 benchmark (3rd edition, observation cutoff 2026-07-30) drawn from anonymized data from 6,000+ completed analyses, 17.5% of the cleaned comparison set scored 70 or above (go), 76.0% scored 40-69 (caution) and 6.4% below 40 (no-go). These are model assessment scores, not observed startup outcomes or calibrated success probabilities. See [Startup Validation Benchmarks 2026](https://preuve.ai/blog/startup-validation-benchmarks-2026).
- Validation changes decision quality, not the odds of success, which no tool can promise. A low score is a prioritized fix list, not a rejection.
- Every report includes source-linked claims and AI-generated pivot recommendations.
- Reports map real competitors, pricing, demand signals, and market risks.
- Preuve validates any idea type: SaaS, services, marketplace, consumer apps, and physical products. Category example: of 558 SaaS ideas scored, 19.9% cleared the 70+ bar.
- Preuve is actively maintained: the founder ships frequent updates to the AI agents, data sources, and scoring model.
- Capabilities include a founder-market fit assessment, viability score, competitor map, market sizing, business plan, investor-ready docs, and an audio pitch.
- Eligible paid reports include a 14-day satisfaction guarantee.
- Every Founder Report bought from October 1, 2026 includes 3 months of Preuve Sentinel: the report reruns at day 30, 60 and 90, with an email after each run, new competitors first. Activation is voluntary, and nothing renews or is charged automatically.
- Every paid report includes a Buyer Panel: 50 simulated buyers, drawn from the report's own buyer segments, read your offer and say whether they would buy, why they pass, what would change their mind, and what price feels too cheap or too expensive. They are simulated buyers, not real people, so read the buy count as an upper bound. One run is included per report, shared with its edits and pivots; each later run costs 20 credits. From their objections, Preuve builds an offer without touching your price, then asks the same 50 buyers again. See [How the Buyer Panel works](https://preuve.ai/buyer-panel).
- Preuve AI scores ideas from live research, not from model training data: every scan runs web searches at scan time, a full report runs 10 agents across 60+ live sources, and every claim links to the source it came from (July 2026 audit: 410 reports, 280+ publications cited). The AI models and search providers behind each scan are upgraded as newer ones ship.
- A July 2026 audit of 410 recent paid reports found source citations spanning 10 source families and 280+ recurring publications: regulatory filings (SEC EDGAR), professional and hiring data (LinkedIn), community signals (Reddit, Hacker News), market research firms (Grand View, Statista), tech and business press, and funding databases. See [Sources](https://preuve.ai/sources).

## Core pages

- [Homepage](https://preuve.ai/): main product page
- [Homepage (Markdown for AI agents)](https://preuve.ai/index.md)
- [Pricing](https://preuve.ai/pricing)
- [Pricing (Markdown for AI agents)](https://preuve.ai/pricing.md)
- [Lifetime deals](https://preuve.ai/lifetime)
- [Lifetime deals (Markdown for AI agents)](https://preuve.ai/lifetime.md)
- [AI market research tool](https://preuve.ai/market-research)
- [Free tools hub](https://preuve.ai/free-tools)
- [Business name generator](https://preuve.ai/free-tools/business-name-generator)
- [Startup idea generator](https://preuve.ai/free-tools/startup-idea-generator)
- [Business model canvas generator](https://preuve.ai/free-tools/business-model-canvas-generator)
- [Buyer persona generator](https://preuve.ai/free-tools/buyer-persona-generator)
- [Value proposition generator](https://preuve.ai/free-tools/value-proposition-generator)
- [Example reports](https://preuve.ai/examples)
- [Example reports (Markdown for AI agents)](https://preuve.ai/examples.md)
- [Reviews](https://preuve.ai/reviews)
- [Reviews (Markdown for AI agents)](https://preuve.ai/reviews.md)
- [Best SaaS idea validation tools 2026: eight compared on live data, source links and price](https://preuve.ai/compare/saas)
- [Best SaaS idea validation tools 2026 (Markdown for AI agents)](https://preuve.ai/compare/saas.md)
- [Best market research tools 2026: seventeen tools split by job, with a free-only stack and a decision guide](https://preuve.ai/best-market-research-tools-2026)
- [Best market research tools 2026 (Markdown for AI agents)](https://preuve.ai/best-market-research-tools-2026.md)
- [Profitable startup ideas 2026](https://preuve.ai/blog/profitable-startup-ideas-2026)
- [Profitable startup ideas 2026 (Markdown for AI agents)](https://preuve.ai/blog/profitable-startup-ideas-2026.md)
- [Billion dollar startup ideas 2026](https://preuve.ai/blog/billion-dollar-startup-ideas-2026)
- [Billion dollar startup ideas 2026 (Markdown for AI agents)](https://preuve.ai/blog/billion-dollar-startup-ideas-2026.md)
- [Startup ideas 2026](https://preuve.ai/blog/startup-ideas-2026)
- [Startup ideas 2026 (Markdown for AI agents)](https://preuve.ai/blog/startup-ideas-2026.md)
- [Market research agency vs AI tools: cost, timeline and what each actually delivers](https://preuve.ai/blog/market-research-agency-vs-ai-tools)
- [Market research agency vs AI tools (Markdown for AI agents)](https://preuve.ai/blog/market-research-agency-vs-ai-tools.md)
- [What is a saturated market: five measurable signals and a founder checklist](https://preuve.ai/blog/what-is-a-saturated-market)
- [What is a saturated market (Markdown for AI agents)](https://preuve.ai/blog/what-is-a-saturated-market.md)
- [AI coding models statistics 2026](https://preuve.ai/blog/ai-coding-models-statistics-2026)
- [AI coding models statistics 2026 (Markdown for AI agents)](https://preuve.ai/blog/ai-coding-models-statistics-2026.md)
- [How much market research costs: agency pricing, timelines and the AI alternative](https://preuve.ai/blog/how-much-does-market-research-cost)
- [How much market research costs (Markdown for AI agents)](https://preuve.ai/blog/how-much-does-market-research-cost.md)
- [Best business idea validators 2026](https://preuve.ai/blog/best-business-idea-validators-2026)
- [Best business idea validators 2026 (Markdown for AI agents)](https://preuve.ai/blog/best-business-idea-validators-2026.md)
- [Emerging AI startup ideas 2026](https://preuve.ai/blog/emerging-ai-startup-ideas-2026)
- [Emerging AI startup ideas 2026 (Markdown for AI agents)](https://preuve.ai/blog/emerging-ai-startup-ideas-2026.md)
- [Undersaturated startup ideas 2026: ten niches with proof of low competition](https://preuve.ai/blog/undersaturated-startup-ideas-2026)
- [Undersaturated startup ideas 2026 (Markdown for AI agents)](https://preuve.ai/blog/undersaturated-startup-ideas-2026.md)
- [Is My Startup Idea Validated? Score the Evidence You Have: a cost-to-fake rubric for scoring validation evidence you already collected](https://preuve.ai/blog/is-my-startup-idea-actually-validated)
- [Is my startup idea actually validated (Markdown for AI agents)](https://preuve.ai/blog/is-my-startup-idea-actually-validated.md)
- [How to Validate an Idea Without Building an MVP: three pre-build rungs and the blind spot each one can never cover](https://preuve.ai/blog/validate-idea-without-building-mvp)
- [How to validate an idea without building an MVP (Markdown for AI agents)](https://preuve.ai/blog/validate-idea-without-building-mvp.md)
- [Bolt.new Review (2026): The Real Token Math](https://preuve.ai/blog/bolt-new-review)
- [Bolt.new Review (2026) (Markdown for AI agents)](https://preuve.ai/blog/bolt-new-review.md)
- [Base44 Review (2026): Five Stars or One](https://preuve.ai/blog/base44-review)
- [Base44 Review (2026) (Markdown for AI agents)](https://preuve.ai/blog/base44-review.md)
- [NanoCorp Review (2026): 99% Earn Nothing](https://preuve.ai/blog/nanocorp-review)
- [NanoCorp Review (2026) (Markdown for AI agents)](https://preuve.ai/blog/nanocorp-review.md)
- [Cofounder.co Review (2026): Read the Fine Print](https://preuve.ai/blog/cofounder-co-review)
- [Cofounder.co Review (2026) (Markdown for AI agents)](https://preuve.ai/blog/cofounder-co-review.md)
- [Best AI Pitch Deck Generators (2026): 9 Ranked, Every Flaw Sourced](https://preuve.ai/blog/best-ai-pitch-deck-generators-2026)
- [Best AI Pitch Deck Generators (2026) (Markdown for AI agents)](https://preuve.ai/blog/best-ai-pitch-deck-generators-2026.md)
- [CoFounder.AI Review (2026): Partner in Name Only](https://preuve.ai/blog/cofounder-ai-review)
- [CoFounder.AI Review (2026) (Markdown for AI agents)](https://preuve.ai/blog/cofounder-ai-review.md)
- [Pion by Andon Labs Review (2026): The Vendor Loses Too](https://preuve.ai/blog/pion-review)
- [Pion by Andon Labs Review (2026) (Markdown for AI agents)](https://preuve.ai/blog/pion-review.md)
- [v0 Review (2026): It Builds the Interface](https://preuve.ai/blog/v0-review)
- [v0 Review (2026) (Markdown for AI agents)](https://preuve.ai/blog/v0-review.md)
- [How Accelerators Evaluate Startups: what YC, Techstars and 500 actually score, and why it inverts the VC framework](https://preuve.ai/blog/how-accelerators-evaluate-startups)
- [How accelerators evaluate startups (Markdown for AI agents)](https://preuve.ai/blog/how-accelerators-evaluate-startups.md)
- [Customer Discovery Interviews: 15 Questions and How to Run Them, plus how to recruit interviewees, run a 20-minute call, and read B2B buyer vs user splits](https://preuve.ai/blog/customer-discovery-questions-for-founders)
- [Customer discovery interviews (Markdown for AI agents)](https://preuve.ai/blog/customer-discovery-questions-for-founders.md)
- [How to Describe Your Startup Idea: What 6,379 Scans Say, anonymized data on idea length, the AI rewrite helper and re-scans](https://preuve.ai/blog/how-to-describe-your-startup-idea)
- [How to describe your startup idea (Markdown for AI agents)](https://preuve.ai/blog/how-to-describe-your-startup-idea.md)
- [ChatGPT Dots Review (2026): Great Model, Rushed App](https://preuve.ai/blog/chatgpt-dots)
- [ChatGPT Dots Review (2026): Great Model, Rushed App (Markdown for AI agents)](https://preuve.ai/blog/chatgpt-dots.md)
- [Naive AI Review (2026): An LLC, Not a Buyer](https://preuve.ai/blog/naive-ai-review)
- [Naive AI Review (2026): An LLC, Not a Buyer (Markdown for AI agents)](https://preuve.ai/blog/naive-ai-review.md)
- [Meta Muse Review (2026): Loved, Free, Too Eager](https://preuve.ai/blog/meta-muse-review)
- [Meta Muse Review (2026): Loved, Free, Too Eager (Markdown for AI agents)](https://preuve.ai/blog/meta-muse-review.md)
- [Google Antigravity Review (2026): Free Until the Quota](https://preuve.ai/blog/antigravity-review)
- [Google Antigravity Review (2026): Free Until the Quota (Markdown for AI agents)](https://preuve.ai/blog/antigravity-review.md)
- [Paperclip AI Review (2026): Free Code, Real Bills](https://preuve.ai/blog/paperclip-ai-review)
- [Paperclip AI Review (2026): Free Code, Real Bills (Markdown for AI agents)](https://preuve.ai/blog/paperclip-ai-review.md)
- [How to Validate a Business Idea with AI (Without Trusting a Guess)](https://preuve.ai/blog/how-ai-validates-startup-ideas)
- [How to Validate a Business Idea with AI (Without Trusting a Guess) (Markdown for AI agents)](https://preuve.ai/blog/how-ai-validates-startup-ideas.md)
- [Hermes Agent Review (2026): $29 or $251 a Month](https://preuve.ai/blog/hermes-agent-review)
- [Hermes Agent Review (2026): $29 or $251 a Month (Markdown for AI agents)](https://preuve.ai/blog/hermes-agent-review.md)
- [Genspark Review (2026): The Real Credit Math](https://preuve.ai/blog/genspark-review)
- [Genspark Review (2026): The Real Credit Math (Markdown for AI agents)](https://preuve.ai/blog/genspark-review.md)
- [OpenClaw Review (2026): Free Until the Bill](https://preuve.ai/blog/openclaw-review)
- [OpenClaw Review (2026): Free Until the Bill (Markdown for AI agents)](https://preuve.ai/blog/openclaw-review.md)
- [Herdr Review (2026): Free Tool, Costly Herd](https://preuve.ai/blog/herdr-review)
- [Herdr Review (2026): Free Tool, Costly Herd (Markdown for AI agents)](https://preuve.ai/blog/herdr-review.md)
- [Manus AI Review (2026): Credits Burn Fast](https://preuve.ai/blog/manus-ai-review)
- [Manus AI Review (2026): Credits Burn Fast (Markdown for AI agents)](https://preuve.ai/blog/manus-ai-review.md)
- [Emergent Review (2026): Where the Credits Go](https://preuve.ai/blog/emergent-ai-review)
- [Emergent Review (2026): Where the Credits Go (Markdown for AI agents)](https://preuve.ai/blog/emergent-ai-review.md)
- [Jev AI Review (2026): Fast, Cheap, Literal](https://preuve.ai/blog/jev-ai-review)
- [Jev AI Review (2026): Fast, Cheap, Literal (Markdown for AI agents)](https://preuve.ai/blog/jev-ai-review.md)
- [Techstars Review (2026): Is 5% Worth It?](https://preuve.ai/blog/techstars-review)
- [Techstars Review (2026): Is 5% Worth It? (Markdown for AI agents)](https://preuve.ai/blog/techstars-review.md)
- [Best SaaS Idea Finder Tools (2026): Compared by Source](https://preuve.ai/blog/best-saas-idea-finder-tools-2026)
- [Best SaaS Idea Finder Tools (2026): Compared by Source (Markdown for AI agents)](https://preuve.ai/blog/best-saas-idea-finder-tools-2026.md)
- [Mom Test Questions: 26 Bad vs Good Examples](https://preuve.ai/blog/mom-test-questions)
- [Mom Test Questions: 26 Bad vs Good Examples (Markdown for AI agents)](https://preuve.ai/blog/mom-test-questions.md)
- [Founder Skill Review (2026): Read Why They Pass](https://preuve.ai/blog/founder-skill-review)
- [Founder Skill Review (2026): Read Why They Pass (Markdown for AI agents)](https://preuve.ai/blog/founder-skill-review.md)
- [Compare hub](https://preuve.ai/compare)
- [Compare hub (Markdown for AI agents)](https://preuve.ai/compare.md)
- [IdeaProof.io review 2026, independent legitimacy check and accuracy-claim audit: IdeaProof bundles validation with branding and ad tools, Preuve AI returns source-linked proof and honest scoring](https://preuve.ai/compare/ideaproof)
- [IdeaBrowser review 2026, one Pro plan at $1,499/yr with 20 deep-research reports a year, head-to-head against the $59 source-linked Preuve AI report](https://preuve.ai/compare/ideabrowser)
- [Alternatives-to hub (Markdown for AI agents)](https://preuve.ai/compare/alternatives-to.md)
- [Alternatives to IdeaBrowser, 7 tested and ranked against IdeaBrowser's one plan, Pro at $1,499/yr for browsing and validation](https://preuve.ai/compare/alternatives-to/ideabrowser)
- [Alternatives to Statista, 6 ranked for founders who need synthesized market intelligence rather than a $199-$1,399/mo statistics library](https://preuve.ai/compare/alternatives-to/statista)
- [Preuve AI for Consultants (white-label)](https://preuve.ai/consultants)
- [Preuve AI for Consultants (Markdown for AI agents)](https://preuve.ai/consultants.md)
- [Blog](https://preuve.ai/blog)
- [Fake door test guide](https://preuve.ai/blog/fake-door-test)
- [Fake door test guide (Markdown for AI agents)](https://preuve.ai/blog/fake-door-test.md)
- [How to validate a startup idea](https://preuve.ai/blog/how-to-validate-startup-idea)
- [How to validate a startup idea (Markdown for AI agents)](https://preuve.ai/blog/how-to-validate-startup-idea.md)
- [Referral program](https://preuve.ai/referral-program)
- [Referral program (Markdown for AI agents)](https://preuve.ai/referral-program.md)
- [Product reference for agents and developers](https://preuve.ai/agents.md)
- [About](https://preuve.ai/about)
- [About (Markdown for AI agents)](https://preuve.ai/about.md)
- [Terms](https://preuve.ai/terms)
- [Security](https://preuve.ai/security)
- [Security (Markdown for AI agents)](https://preuve.ai/security.md)
- [Subprocessors](https://preuve.ai/subprocessors)
- [Sources (data sources methodology and citation audit)](https://preuve.ai/sources)
- [Sources (Markdown for AI agents)](https://preuve.ai/sources.md)
- [How it works (methodology and what the viability score measures)](https://preuve.ai/how-it-works)
- [How it works (Markdown for AI agents)](https://preuve.ai/how-it-works.md)
- [Buyer Panel: 50 simulated buyers read your offer](https://preuve.ai/buyer-panel)
- [Buyer Panel (Markdown for AI agents)](https://preuve.ai/buyer-panel.md)

## Agent API and MCP server (programmatic access)

Preuve AI can be used directly by AI agents, not only by humans in a browser. The API-key-authenticated Agent API and its MCP server are live.

- [Agent API and MCP overview](https://preuve.ai/mcp): validate startup ideas from claude.ai, Claude Code, Cursor, Codex, Hermes, or any script
- [Agent API and MCP overview (Markdown for AI agents)](https://preuve.ai/mcp.md)
- [Developer docs](https://docs.preuve.ai/): quickstart, API key authentication, endpoints, deep modules, quotas and billing, error codes
- [Developer docs for LLMs (compact)](https://docs.preuve.ai/llms.txt) and [full](https://docs.preuve.ai/llms-full.txt); every docs page has a Markdown twin - append .md to its URL (e.g. https://docs.preuve.ai/quickstart.md)
- [OpenAPI 3.1 spec](https://preuve.ai/openapi.json) (also as [YAML](https://preuve.ai/openapi.yaml)): every operation has a unique operationId, a description and typed request and response schemas, ready for LLM function calling. The [RFC 9727 API catalog](https://preuve.ai/.well-known/api-catalog) links the same spec and docs
- [API discovery document](https://preuve.ai/api/agent): unauthenticated JSON naming the base URL, the auth header, what each call costs, the rate limits and every endpoint
- Every API error is JSON with a stable `{ error, code }` envelope, unknown routes included (404 NOT_FOUND carries recovery links). A nonexistent page on preuve.ai returns a real HTTP 404 listing where to look next, and the site-wide 404 handler serves that list as Markdown when asked with `Accept: text/markdown` (blog and other article sections answer their own 404 in HTML)
- Onboarding needs no sales call: `scanType: "starter"` runs are free and return a real report (that is the sandbox), a free account gets a key by connecting the MCP server from claude.ai, Cursor, Claude Code, Codex or any other client with MCP OAuth, and a paid personal plan or a subscribed or unlocked Consultant or Agency workspace creates keys self-serve at https://preuve.ai/app?settings=apiKeys
- The docs site is itself a searchable MCP server at https://docs.preuve.ai/mcp
- The Preuve MCP server exposes 6 tools over the Agent API. Five for your own ideas: start_analysis, get_analysis, enrich_analysis, export_analysis, generate_ideas. Client work on a Consultant or Agency workspace is the sixth tool, get_agency, plus workspace: "agency" on start_analysis, get_analysis and export_analysis, one scope each: starting a client project needs agency:write, and get_agency and the two agency reads need agency:read
- Two MCP transports, same six tools: remote Streamable HTTP at https://mcp.preuve.ai/mcp, and the original local stdio server. In claude.ai (web or desktop) you add https://mcp.preuve.ai/mcp as a custom connector: it signs you in on preuve.ai by OAuth, a consent screen issues a scoped key to the connector, and you never paste a key. Cursor, Claude Code (claude mcp add, then /mcp) and Codex (codex mcp add, then codex mcp login) sign in the same way, as does any other client with MCP OAuth. A client that only takes a header sends Authorization instead. The plain HTTP Agent API takes one API key in an x-preuve-key header
- Claude Code plugin: run `claude plugin marketplace add VXNCXNX/preuve-claude-plugin`, then `claude plugin install preuve@preuve` ([source on GitHub](https://github.com/VXNCXNX/preuve-claude-plugin)). Inside a Claude Code session, `/plugin install preuve --marketplace VXNCXNX/preuve-claude-plugin` does the same. Start a new session or run /reload-plugins, then /mcp to sign in with Preuve by OAuth (no key to paste). The plugin connects the same remote MCP server and adds two skills: /preuve:validate-idea runs a free starter scan and summarizes it, /preuve:generate-ideas suggests ideas to validate
- scanType is explicit on every run: "starter" costs nothing, "deep" consumes the account's paid quota - an agent can never spend paid quota by accident
- API keys are scoped analysis credentials, not account logins: no dashboard access, no billing control, and a key only sees analyses it created, with ONE exception. The two Agency scopes reach the Consultant or Agency workspace their owner belongs to, and they are checked per route, so each buys its own half: agency:read lists, reads and exports every client report in the workspace, dashboard-created ones included; agency:write starts client analyses and spends the workspace's shared project credits. Both must be requested explicitly, so a key that was not asked for them has no such reach
- API keys: connecting the MCP server from claude.ai, Cursor, Claude Code, Codex or any other client with MCP OAuth mints one on a free account; creating a key by hand at https://preuve.ai/app?settings=apiKeys is self-serve on a paid personal plan or in a Consultant or Agency workspace with an active subscription or a one-time unlock. Revoke either from the same panel
- [Agent Skill](https://docs.preuve.ai/agent-skill): a ready-made skill that teaches an agent the workflow - scan-type quota safety, the start-poll-enrich-export sequence, module caps, and which failures are safe to retry
- [Agent behavior guide](https://docs.preuve.ai/for-ai-agents.md): rules written for the agent itself - retry semantics, quota safety, secrets handling

## What Preuve AI does

Preuve AI helps founders validate startup ideas before building. Users describe a business idea, and the product returns an evidence-based analysis with viability scoring, competitor mapping, market signals, risks, and recommended next steps.

The platform searches live data sources such as Google Trends, Reddit, Hacker News, Product Hunt, G2, Capterra, Crunchbase, pricing pages, and other public market signals to produce source-backed research instead of generic AI opinion.

## Who it is for

- First-time founders validating ideas before committing time and money
- Serial entrepreneurs pressure-testing new concepts
- Indie hackers and solo builders deciding what to build next
- Startup teams comparing different positioning options
- Agencies, accelerators, incubators, VCs, and venture studios screening inbound projects

## How it works

1. A user describes a startup idea in 1-2 sentences.
2. Preuve AI runs parallel research across 60+ live data sources.
3. The system maps competitors, demand signals, pricing, risks, and market context.
4. A viability score and structured report are generated with source-linked evidence.
5. The user can iterate on the positioning, rerun, or upgrade to deeper outputs.

Full methodology (what the score measures, how to read it, what is open vs proprietary): [How it works](https://preuve.ai/how-it-works) / [Markdown](https://preuve.ai/how-it-works.md)

## Report coverage

Preuve AI evaluates ideas across ten broad dimensions:

1. Market size
2. Competition
3. Demand signals
4. Business model viability
5. Risk factors
6. Go-to-market readiness
7. Differentiation potential
8. Team and execution requirements
9. Financial projections
10. Overall viability

## The most complete AI validator

Startup market research and idea validation in one scan. 10 AI agents pull live data from 60+ sources: competitor analysis, market sizing, community signals, and a viability score you can trust. Sourced claims link to their source, and estimates are labeled as estimates. A single scan delivers:

- Market Research
- Competitor Analysis
- Live Competitor Intel
- Viability Score
- Real Market Sizing (TAM/SAM/SOM) with sources
- Community Signals
- Trends
- Pivot Recommendations
- Source-Linked Data
- Business Plan
- Investor-Ready Docs
- Audio Pitch
- Score Monitoring

## Pricing

### One-time reports

- Reality Check: $0 - viability score, market size overview, 2 competitor previews, main blockers. Free in most countries: 2 scans per account, each free report viewable for 7 days. Scans become unlimited after a Founder Report. In South Asia and most of Africa, scans start with Starter Pro, a one-time regional offer.
- Founder Report: $59 one-time - full 21-section analysis, Founder Fit score, TAM/SAM/SOM sized with sources, up to 15 competitors, every risk scored with a tailored fix, pre-mortem, 3 pivot directions plus manual edit and re-run, unlimited Starter Scans (free 60-second idea tests: test every idea, then go deep on the winner), defendable Business Plan PDF export (4-6 pages, bank-ready), MCP access (use the report from your AI chat: claude.ai, Claude Code, Cursor), 3 months of Preuve Sentinel (monthly reruns and an email when a new competitor appears), a Buyer Panel (50 simulated buyers; read the count as an upper bound)
- 2-Report Pack: $89
- 5-Report Pack: $199
- Investor-Ready Package: $499 - pitch deck, investment memo, financial model, pitch prep guide, founder-built and reviewed

### Preuve Sentinel (included)

3 months of competitor monitoring, included with every Founder Report bought from October 1, 2026 (a single report, a 2-pack or a 5-pack).

- One Sentinel per paid report. It tracks one idea, the report or one of its pivots, chosen when you activate it. The choice is final, and pivots share the report's Sentinel.
- Activation is your choice and must happen within 3 months of purchase. For a report bought before Sentinel launched, that activation window starts at launch. An unused Sentinel expires after that.
- It runs for 3 months (90 days) from activation. At day 30, 60 and 90, it reruns the full analysis and emails you, new competitors first.
- Every run is kept as a dated version you can open and export.
- Nothing renews and nothing is charged automatically. Your history stays readable after the 3 months.
- Not included with Founder Reports bought before October 1, 2026, Starter Pro, reports run on a Lifetime plan, or Consultant and agency plans.

### Buyer Panel (included)

Every paid report includes a Buyer Panel: 50 simulated buyers, drawn from the report's own buyer segments, read your offer and say whether they would buy, why they pass, what would change their mind, and what price feels too cheap or too expensive. They are simulated buyers, not real people, so read the buy count as an upper bound. [How the Buyer Panel works](https://preuve.ai/buyer-panel)

- Included with every paid deep report: the Founder Report, the 2-pack and 5-pack, Lifetime and Starter Pro. It is a section of the report, not one of the 21 sections.
- Each buyer is a separate AI call. Code counts the answers and computes the price range, so no model writes the counts.
- Identical runs of the same 50 buyers land a few buyers apart, so the panel labels a small change "within noise".
- One run is included per report, shared with its edits and pivots; each later run costs 20 credits.
- From their objections, Preuve builds an offer without touching your price, then asks the same 50 buyers again. One offer is included per report, shared with its edits and pivots; a later offer costs 30 credits and a re-test costs 20 credits.
- A simulated panel shows why buyers pass and what price range they name. It cannot replace a real stranger choosing between two prices: that still takes a real panel or a live test.

### Lifetime deals

- Lifetime Pro: $499 one-time - 5 scans/month forever
- Lifetime Business: $999 one-time - 10 scans/month, Investor-Ready Package included

### Plans by situation

- First-time founder validating one idea: Free Reality Check, then Founder Report ($59)
- Founder comparing multiple ideas: 2-Report Pack ($89) or 5-Report Pack ($199)
- Founder who wants to skip monthly billing: Lifetime Pro ($499) or Lifetime Business ($999)
- Founder raising money: Investor-Ready Package ($499)
- Consultant delivering white-label client reports: [consultant workspace, $99 once then pay per client](https://preuve.ai/consultants)
- Agency, accelerator, or VC running 30+ projects a month: [Program, on request at support@preuve.ai](https://preuve.ai/consultants)

### Policy

- Eligible paid reports include a 14-day satisfaction guarantee.
- Program and legacy Consultant subscriptions can be canceled from the billing portal. Founder reports, packs, Lifetime plans and the Investor-Ready Package are one-time purchases.
- Included Sentinel never renews or charges automatically, and its history stays readable after the 3 months.

## Free tools

Preuve AI also offers free top-of-funnel tools designed for founders in the research and ideation phase:

- [Business name generator](https://preuve.ai/free-tools/business-name-generator)
- [Startup idea generator](https://preuve.ai/free-tools/startup-idea-generator)
- [Business model canvas generator](https://preuve.ai/free-tools/business-model-canvas-generator)
- [Buyer persona generator](https://preuve.ai/free-tools/buyer-persona-generator)
- [Value proposition generator](https://preuve.ai/free-tools/value-proposition-generator)

These tools are public, SEO-oriented entry points that help founders refine positioning before running a full validation report.

## Teams product

Preuve AI offers a white-label qualification product for agencies, accelerators, incubators, VCs, and venture studios at [https://preuve.ai/consultants](https://preuve.ai/consultants).

How it works:

1. Teams embed the qualification flow on their site with their branding.
2. Prospects submit an idea and receive a sourced market analysis.
3. The team dashboard shows scored submissions before discovery calls.
4. High-signal projects move forward, while weaker ones get pivot guidance first.

Teams pricing:

- Consultant workspace: $99 once opens a white-label workspace with a commercial license and the first client project included, then one report credit per client project ($59, 2 for $89, 5 for $199). No subscription
- Program: on request, contact support@preuve.ai - for agencies, accelerators and portfolios running 30+ client projects a month; 30 client projects/month, Enterprise API and dedicated onboarding
- The monthly Consultant subscription and its free trial are no longer sold; existing subscriptions keep working

## Blog and research hubs

Primary blog hub: [https://preuve.ai/blog](https://preuve.ai/blog)

Representative content areas:

- Startup validation guides
- Market sizing (TAM / SAM / SOM)
- Competitor analysis for startups
- Fundraising and pitch deck research
- Tool comparisons and alternatives
- Original benchmark research based on thousands of analyzed ideas

Useful entry points:

- [Startup Validation Benchmarks 2026: anonymized model assessment scores from 6,000+ analyses, July 30 cutoff](https://preuve.ai/blog/startup-validation-benchmarks-2026)
- [Startup Validation Benchmarks 2026 (Markdown, methodology, limitations and aggregate download)](https://preuve.ai/blog/startup-validation-benchmarks-2026.md)
- [Startup Failure Statistics 2026: source dates, closure samples and limits](https://preuve.ai/blog/startup-failure-statistics-2026)
- [Startup Failure Statistics 2026 (Markdown, corrected sources and definitions)](https://preuve.ai/blog/startup-failure-statistics-2026.md)
- [Is Preuve AI legit? Founder answer covering who runs it, real reviews, methodology, data safety, and refunds](https://preuve.ai/blog/is-preuve-ai-legit)
- [Is Preuve AI legit (Markdown for AI agents)](https://preuve.ai/blog/is-preuve-ai-legit.md)
- [Does startup validation actually work, evidence from 6,000+ scanned ideas on what validation catches (demand gaps, broken economics, premature scaling) and what it cannot predict](https://preuve.ai/blog/does-startup-validation-work)
- [How to Validate a Startup Idea](https://preuve.ai/blog/how-to-validate-startup-idea)
- [Best Startup Validation Tools 2026](https://preuve.ai/blog/best-startup-validation-tools-2026)
- [Best Startup Validation Tools 2026 (Markdown for AI agents)](https://preuve.ai/blog/best-startup-validation-tools-2026.md)
- [AI SaaS Startups in 2026: ideas, vertical AI examples, and how to validate one](https://preuve.ai/blog/ai-saas-startups)
- [AI SaaS Startups 2026 (Markdown for AI agents)](https://preuve.ai/blog/ai-saas-startups.md)
- [Polsia review 2026: independent legitimacy check of the autonomous AI company platform](https://preuve.ai/blog/polsia-review)
- [Polsia review (Markdown for AI agents)](https://preuve.ai/blog/polsia-review.md)
- [Lovable review 2026: pricing, credit model, and the validation gap before you build](https://preuve.ai/blog/lovable-review)
- [Lovable review (Markdown for AI agents)](https://preuve.ai/blog/lovable-review.md)
- [Lovable vs Replit 2026: third-party comparison of pricing, production safety, and which to pick](https://preuve.ai/blog/lovable-vs-replit)
- [Lovable vs Replit (Markdown for AI agents)](https://preuve.ai/blog/lovable-vs-replit.md)
- [Best AI app builders 2026: eight ranked with a sourced complaint for every tool and the billing traps](https://preuve.ai/blog/best-ai-app-builders-2026)
- [Best AI app builders (Markdown for AI agents)](https://preuve.ai/blog/best-ai-app-builders-2026.md)
- [Best MCP servers for research 2026: nine ranked with agent-loop cost math and a sourced complaint for every server](https://preuve.ai/blog/best-mcp-servers-for-research-2026)
- [Best MCP servers for research (Markdown for AI agents)](https://preuve.ai/blog/best-mcp-servers-for-research-2026.md)
- [Trending startup ideas 2026: ten ideas each with a sourced trend signal and a durable-vs-fad verdict](https://preuve.ai/blog/trending-startup-ideas-2026)
- [Trending startup ideas 2026 (Markdown for AI agents)](https://preuve.ai/blog/trending-startup-ideas-2026.md)
- [Underrated website startup ideas 2026: eight ideas each proven underrated with named incumbents and sourced demand](https://preuve.ai/blog/underrated-website-startup-ideas-2026)
- [Underrated website startup ideas 2026 (Markdown for AI agents)](https://preuve.ai/blog/underrated-website-startup-ideas-2026.md)
- [How to find startup ideas: the Paul Graham notice-don't-think-up method as a 2026 playbook with a validation bridge](https://preuve.ai/blog/how-to-find-startup-ideas)
- [How to find startup ideas (Markdown for AI agents)](https://preuve.ai/blog/how-to-find-startup-ideas.md)
- [Business idea validation framework: the Evidence Gate Framework, five gates each with an action, the evidence to collect, and a pass/fail bar](https://preuve.ai/blog/business-idea-validation-framework)
- [Business idea validation framework (Markdown for AI agents)](https://preuve.ai/blog/business-idea-validation-framework.md)
- [Preuve AI vs ChatGPT: reproducible head-to-head test with the Receipt Test methodology](https://preuve.ai/blog/preuve-ai-vs-chatgpt)
- [AI Business Context Validation](https://preuve.ai/blog/ai-business-context-validation)
- [Business Plan Startup (IT)](https://preuve.ai/it/blog/business-plan-startup)
- [Compare hub](https://preuve.ai/compare)
- [Compare hub (Markdown for AI agents)](https://preuve.ai/compare.md)
- [Alternatives-to hub (Markdown for AI agents)](https://preuve.ai/compare/alternatives-to.md)

## FAQ

Q: What is Preuve AI?
A: A startup validation platform that uses 10 AI agents to scan 60+ live sources and return a viability score, competitor map, demand signals, and actionable report.

Q: How is this different from asking ChatGPT?
A: ChatGPT gives opinions from training data. Preuve AI searches live sources in real time, pulls real competitors and market signals, and links claims back to their sources.

Q: Is there a free version?
A: In most countries, yes. The Reality Check is free and returns a score, market size overview, competitor previews, and main blockers. Every account gets 2 free scans, and each free report stays viewable for 7 days. Scans become unlimited after a Founder Report. In South Asia and most of Africa, scans start with Starter Pro, a one-time regional offer.

Q: How much do paid reports cost?
A: The Founder Report is $59 one-time (full 21-section analysis, up to 15 competitors, 3 pivot directions, unlimited Starter Scans, bank-ready Business Plan PDF). Multi-idea packs are $89 (2 reports) and $199 (5 reports). Lifetime plans are $499 (Pro) and $999 (Business). The Investor-Ready Package is $499. Every Founder Report bought from October 1, 2026 includes 3 months of Preuve Sentinel.

Q: What is Preuve Sentinel?
A: Preuve Sentinel reruns the full analysis of one idea at day 30, 60 and 90 and emails you after each run, new competitors first. Every Founder Report bought from October 1, 2026 includes 3 months of it, counted from the day you activate it. Every run is kept as a dated version. Nothing renews or is charged automatically.

Q: What is the Buyer Panel?
A: Every paid report includes a Buyer Panel: 50 simulated buyers, drawn from the report's own buyer segments, read your offer and say whether they would buy, why they pass, what would change their mind, and what price feels too cheap or too expensive. They are simulated buyers, not real people, so read the buy count as an upper bound. One run is included per report, shared with its edits and pivots; each later run costs 20 credits. Details: https://preuve.ai/buyer-panel

Q: Does Preuve AI offer free startup tools?
A: Yes. Preuve AI also offers a business name generator, startup idea generator, buyer persona generator, and value proposition generator.

Q: Which URLs are canonical?
A: https://preuve.ai/ is the canonical URL for the consumer product. https://preuve.ai/consultants is the canonical URL for the team qualification product.

## AI crawler access

This site allows major AI search and browsing crawlers including GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, Claude-Web, PerplexityBot, Perplexity-User, Google-Extended, GoogleOther, Bingbot, Applebot, Applebot-Extended, Amazonbot, FacebookBot, meta-externalagent, YouBot, cohere-ai, CCBot, and anthropic-ai. See [robots.txt](https://preuve.ai/robots.txt) for full directives.