Skip to content

FAQ

LMIF (Look Ma I’m Famous) is an open protocol for identity protection in the AI era. It allows individuals and organizations to claim and control how their likeness is used in AI applications.

How is LMIF different from DMCA takedowns?

Section titled “How is LMIF different from DMCA takedowns?”

DMCA is reactive - you find violations and request takedowns. LMIF is proactive - platforms check before allowing content creation. This prevents violations rather than just responding to them.

Any platform that creates or hosts AI avatars, chatbots, or other content that might use someone’s likeness. This includes:

  • AI companion platforms
  • Social media with AI features
  • Gaming companies
  • Content creation tools
  1. Sign up at lookmaimfamous.com/developers
  2. Get your API key
  3. Follow the Quick Start guide
  4. Implement identity checks in your avatar creation flow

Yes. Every avatar that might use a real person’s likeness should be checked. This includes:

  • Celebrity-inspired avatars
  • User-uploaded photos
  • Text descriptions of real people

Implement appropriate fallback behavior:

  • Fail-closed (recommended): Block creation if you can’t verify
  • Fail-open (risky): Allow creation but mark for later review
  • Layer 1 (exact match): under 100ms
  • Layer 2 (AI analysis): 2-5 seconds
  • Layer 3 (human review): 24-48 hours

Most checks complete in Layer 1 and return near-instantly.

LMIF uses a 3-layer system:

  1. Registry match: Fast database lookup
  2. AI analysis: Vision and semantic analysis for variations
  3. Human review: Edge cases and appeals

The AI analysis layer catches common evasion attempts like:

  • Leetspeak (T4ylor Sw1ft)
  • Phonetic spellings (Tay Tay)
  • Visual similarity without exact name
  • Deepfakes

No system is 100% perfect, but compliance platforms are expected to make good-faith efforts.

LMIF includes parody detection. Parody content may be allowed depending on the identity owner’s policy. If detected as parody, the API returns REVIEW_PARODY action.

Requiring both prevents false positives. “John Smith” is a common name - we can’t block every John Smith. But “John Smith” + a celebrity’s face = clear match.

PolicyMeaning
BLOCK_ALLNo AI avatars allowed
BLOCK_COMMERCIALPersonal use OK, commercial blocked
MONETIZEAllowed with revenue share
LICENSERequires explicit license
TEAMOnly authorized accounts
OPENAnyone can create (monitoring only)

Yes. Identity owners can change their policy at any time. If a policy becomes more restrictive, existing avatars enter a grace period.

A 30-day window for existing avatars to comply when an identity is boxed or policy changes. During this time:

  • Avatars continue working
  • Creators are notified
  • Users are warned
  • Options provided (license, modify, remove)
  1. Check the identity - if policy is LICENSE, you’ll get license tier information
  2. Request a license through the API
  3. Identity owner approves/denies
  4. If approved, you receive a license ID

Prices are set by identity owners. Typical ranges:

  • Personal: $0-10/month
  • Creator: $25-100/month
  • Commercial: $100-1,000/month
  • Enterprise: Custom

You cannot create avatars of that identity. Consider:

  • Modifying your use case
  • Creating original characters
  • Appealing with additional information

For MONETIZE policy:

  1. Create avatar with tracking enabled
  2. Track usage (messages, subscriptions, tips)
  3. Report usage monthly
  4. Royalties calculated and paid to identity owner

Set by identity owners, typically 5-20% of avatar revenue. The exact rate is returned in the API response.

Monthly, by the 15th of the following month. Late reports may incur penalties.

At minimum:

  • box.created - Scan for existing violations
  • grace_period.started - Notify affected users
  • grace_period.expired - Enforce compliance

Webhooks are retried for up to 24 hours. You can also poll the API for grace periods affecting your platform.

Webhooks may be delivered multiple times. Implement idempotent handlers using the event ID.

Basic: Identity check before avatar creation Standard: + Webhooks, grace periods, user migration Premium: + Licensing, revenue sharing, audits

Non-compliant platforms may:

  • Lose access to licensed identities
  • Face legal action from identity owners
  • Be publicly flagged as non-compliant
  1. Complete self-assessment
  2. Demonstrate integration in sandbox
  3. Request compliance review
  4. Pass review
  5. Receive badge

JPEG, PNG, WebP, GIF (first frame). Max size: 10MB.

No. The SDK is server-side only. Never expose your API key in client-side code.

Not currently. The REST API is our primary interface.

Currently only TypeScript. Python and Go SDKs are planned.

Email developers@lookmaimfamous.com with:

  • Request ID from error response
  • Steps to reproduce
  • Expected vs actual behavior

Yes: status.lookmaimfamous.com