← Idea board

Finance & compliance

Ephemeral ID

Zero-retention age and identity checks for small apps and EU SMBs that need a signed attestation, not stored ID documents.

Founder thesis

Small apps that need age or ID checks sit between $49-$250/month vendor minimums and vendors that keep the documents. The trust gap is public: retention, forced verification, scope creep, no deletion guarantee. Bet: a zero-retention, self-hostable attestation service takes that segment on price and trust, and new age-assurance mandates keep pushing platforms to third-party vendors.

What others got wrong

  • Persona gates low volume behind a $250/month Essential minimum on a 12-month term, so a small app pays for capacity it does not use.

  • Veriff's cheapest self-serve plan carries a $49/month minimum and charges an extra $0.30 for extended retention.

  • Users are left relying on internal policies, not a legal guarantee, for how long their biometrics are kept.

  • Identity checks are injected into paid upgrade flows and block existing access when refused.

What should work

  • Publish every module at a per-success price with free checks and no minimums, the move Didit made with 500 free checks/month and $0.33 per full KYC.

  • Ship an agent-native funnel: a published integration prompt, MCP server and agent skills so an AI coding agent provisions a live flow from one prompt.

  • Return a signed attestation only, delete documents and selfies at the edge, and offer an optional self-hosted NFC e-passport and on-device-liveness path for the most privacy-sensitive customers.

  • Flat $0.40-$0.60 per check with no monthly minimum, below Persona and Veriff and above Didit's $0.30 module cost base.

What kills it

  • The retention complaints are secondhand HN comments, one a pasted ChatGPT answer.
  • Licensing and per-country data sources make this capital-intensive, not an indie play.
  • Willingness to pay for zero-retention is unproven.
  • A funded incumbent could copy a no-retention policy with one feature launch.

What users want

5
  • A clear answer to what was kept and who got it after a government-ID-plus-selfie check.

    One HN commenter pasted a ChatGPT answer claiming Persona keeps biometric data up to one year, with no legal guarantee.

  • No identity check injected into a paid upgrade flow.

    An HN commenter wrote that a Persona check appeared mid-upgrade and declining it blocked all access to their original Claude Code plan.

  • Deletion that is guaranteed, not open-ended retention.

    An HN commenter restated Veriff's privacy notice, which permits keeping data whenever it is necessary for services or legal claims.

  • A verification that does not silently expand beyond what was consented to.

    An HN commenter said a researcher found an exposed Persona frontend running 260+ checks, including facial recognition and adverse media screening.

  • An appeal path instead of a dead end after a false reject.

    An HN commenter said a 15-year-old paid LinkedIn account was asked for facial scan, REAL ID and passport, with no one to contact.

Sources
11
Checked
11
Competitors
3
Gaps
6
Source depth
Page checked
11 · 100%
Snippet only
0 · 0%
Claim confidence
Documented
7 · 54%
Interpreted
6 · 46%
Sources by type
  • complaint5
  • pricing3
  • revenue1
  • market1
  • playbook1

Competitors

3
Starting monthly price

Parsed from public pricing blurbs (entry paid tier when available). Max in set: $250/mo.

Competitor comparison
CompetitorFromFree tierSource
Persona$250/moYeswithpersona.com✓
Veriff$49/mo—veriff.com✓
DiditFreeYesdidit.me✓

What customers complain about

6
What customers complain about
GapSrc
An HN commenter claims Persona keeps biometric data for up to a year with no legal guarantee, but the text is a ChatGPT answer they pasted, so the retention term is unverified.
A paying Claude Code user met a Persona check mid-upgrade, and declining blocked access to their original plan.
Veriff's retention is reported as open-ended, but that comes from an HN commenter restating the privacy notice, not the page itself.
An HN commenter says an exposed Persona frontend ran 260+ checks, but the comment links no primary research, so it is secondhand.
A long-standing paid LinkedIn account was flagged and asked for facial scan, REAL ID and passport with no appeal.
Low-volume or seasonal apps needing a few hundred checks pay for unused capacity because Persona's Essential is $250/month on a 12-month minimum and Veriff's cheapest self-serve plan carries a $49/month minimum.

Documented Interpreted

How competitors find customers

3
How competitors got customers
TacticSrc
Persona's startup program gives early-stage companies up to 12 months free (500 Government ID verifications and 1,000 Relay services/month, $1 after in the General cohort) through accelerator and community partners, then converts them to the $250/month Essential plan.withpersona.com✓
Didit runs programmatic comparison and alternative pages (vs Persona, Veriff, Sumsub, Onfido, Jumio, Stripe Identity) plus an 'Ask an AI to summarise this page' block that seeds ChatGPT, Claude and Perplexity answers, tied to a public per-module price list.didit.me✓
Didit's agent-first funnel publishes an integration prompt, MCP server and agent skills so an AI coding agent can integrate end to end and generate keys unattended.didit.me✓

Your angle & pricing wedge

Angle and pricing wedge
PieceClaimSrc
AngleVerify once and return only a short-lived signed attestation (over-18, identity-verified, unique-human plus a salted per-service pseudonym). Raw documents and selfies are deleted at the edge and never written to durable storage, with an optional self-hosted, open-source NFC e-passport and on-device-liveness path for the most privacy-sensitive customers.
WedgeFlat per-verification price, target $0.40-$0.60 with no monthly minimum, which undercuts Persona's $250/month Essential minimum on a 12-month term and Veriff's $49-$209/month minimums while matching Didit's transparency at $0.33 full KYC and 500 free checks/month. The mechanism is unbundling verification compute from long-term storage: the marginal cost of a check is document/OCR plus liveness plus face-match compute, which Didit prices at $0.15 + $0.10 + $0.05 = $0.30, so a zero-retention vendor does not carry storage, manual-review or retention-add-on costs like Veriff's +$0.30 for extended retention.

How others got traction

3
What worked for others
MoveWhySrc
Didit published every verification module at a per-success price with 500 free checks/month and no minimums, instead of quote-only enterprise pricing.Its founder says opaque pricing hidden behind sales calls is what it flipped, and that this transparency produced >30% MoM growth and 2,000+ companies; removing the sales gate is the likely reason developers and agents can start without a contract.didit.me✓
Agent-native onboarding: a published integration prompt, MCP server and agent skills let an AI coding agent provision a live verification flow from one prompt.Didit says it doubled down on developers and AI agents, and the likely effect is that agent skills make the vendor recommendable inside ChatGPT, Claude and Perplexity answers, a new discovery channel.didit.me✓
Persona's startup program gives up to 12 months free with 500 Government ID verifications and 1,000 Relay services/month, $1 thereafter in the General cohort, distributed through accelerator and community partners.The page frames it as land-and-expand into hundreds of early-stage startups that later convert to the $250/month Essential plan; the likely effect is that partner channels pre-qualify early fintechs and lower acquisition cost.withpersona.com✓

Market heat

  • Didit became profitable and grew >30% month-over-month to 2,000+ companies after a $6M seed, total $7.5M, and reports 80% of customers had never used an identity-verification provider before, self-reported.

    More ▾
  • Persona's self-serve Essential plan starts at $250/month with a minimum 12-month contract and pay-per-successful-verification billing.

  • Didit's competitor comparison estimates Persona at a $500/month minimum and $1.50 per KYC flow, marked estimated from G2, Capterra, Vendr and reseller listings.

    More ▾
  • Age-assurance and identity-verification mandates, such as Australia's under-16 social-media law enforced from 10 Dec 2025, push platforms to third-party ID vendors; one HN commenter on that law alleges a Persona frontend performing 260+ checks, which may exceed what users consented to.

    More ▾
  • Public backlash against retention-heavy ID vendors shows up across Claude (forced Persona verification mid-upgrade), LinkedIn false rejects and Porkbun/Veriff retention.

    More ▾

Revenue signals

5
Revenue signals
SignalSrc
Didit, self-reported: $7.5M total raised, profitable, >30% MoM growth, 2,000+ companies, 80% of customers new to identity verification.
didit.me✓
Persona public list price: Essential from $250/month with a minimum 12-month contract, pay per successful verification.
withpersona.com✓
Veriff public list prices: $0.80-$1.89 per verification with $49-$209/month minimums.
veriff.com✓
Didit public list price: $0.33 per full KYC with no minimum and 500 free verifications/month.
didit.me✓
A privacy-first entrant could charge roughly $0.40-$0.60 per verification with no minimum and still sit below Persona and Veriff while above Didit's published module cost base. Arithmetic: 1,000 customers x 1,500 checks/month = 1.5M checks/month; at $0.40-$0.60 that is roughly $0.6M-$0.9M/month, or about $7.2M-$10.8M ARR. Willingness to pay is unproven.

Should you build this?

This is not a low-capital indie play: Didit is profitable at >30% MoM self-reported, and Persona and Veriff sell against annual minimums. Didit says the incumbent moat is licenses and per-country data sources, the hard, unglamorous work. The wedge is the documented trust gap, not price alone: HN complaints target mandatory retention, alleged 260-check scope creep, forced verification mid-upgrade and no deletion guarantees. But the complaint evidence is secondhand and one item is AI-generated, and licensing may force a narrower age-only or e-passport-only scope. Test first: publish a one-page zero-retention age-verification spec plus a waitlist, cold-email 50 operators of privacy-sensitive communities and EU SMBs facing age-assurance duties, and ask for a refundable $99 pre-order for 1,000 checks; proceed only if at least 10 pay.

More ▾

All sources

11