PaperVideo — Privacy Policy

• Customer — the individual or entity that uploads a research paper and orders a PaperVideo.
• Presenter — an individual whose photographic likeness and/or voice clone is used to generate an AI avatar presenter in a PaperVideo.
• Biometric Data — data derived from an individual's physical characteristics used for identification or avatar generation purposes, including facial geometry extracted from photographs and voiceprint characteristics extracted from audio recordings.
• Presenter Assets — photographs, audio recordings, voice samples, and any other materials provided by or on behalf of a Presenter for the purpose of generating AI avatar videos and synthesized voiceovers.
• Generated Content — the AI-generated video, avatar animations, synthesized voiceover audio, and animated HTML scenes produced by the PaperVideo pipeline.
• Third-Party AI Providers — external service providers we use to process Presenter Assets and generate components of PaperVideo videos, currently including HeyGen (avatar generation), ElevenLabs (voice synthesis), and Kling/Replicate (video rendering).

2.1 Information You Provide as a Customer

• Account information (name, email, institutional affiliation)
• Research papers (PDFs) you upload for video conversion
• Script review edits and customization preferences (avatar placement, wardrobe selection, language preferences)
• Payment and billing information processed through Stripe or Apple (we do not store full payment card numbers)
• Institutional or sponsor logo files uploaded for overlay
• Communications with us regarding your orders

2.2 Information We Collect from Presenters

When an individual consents to serve as a Presenter, we collect:

• Photographs — up to three (3) reference photographs (minimum one), showing the Presenter in upper-body, waist-up, frontal-view framing
• Voice recordings — audio samples or recordings used to create a voice clone via ElevenLabs
• Biometric Data — facial geometry and voiceprint characteristics derived from the above materials by our Third-Party AI Providers, used solely for generating avatar animations and synthesized speech
• Identity information — full name, professional title, institutional affiliation, and any additional credentials the Presenter wishes to display
• Signed consent — a completed Presenter Consent and Likeness Release form

2.3 Information Extracted from Research Papers

Our automated pipeline uses AI (Claude API by Anthropic) to extract from uploaded PDFs:

• Author names, affiliations, and ORCID identifiers
• Study metadata (study type, journal name, publication date, DOI)
• Statistical data, results, and findings as reported in the paper
• Clinical context, background information, and discussion points

This extraction is performed solely to generate the video content and is not used for any other purpose.

2.4 Information Collected Automatically

• Usage data: features accessed, videos ordered, revision requests, delivery preferences, language variant selections
• Device and browser information
• Log data: IP address, access times, error reports

2.5 Information We Do NOT Collect

• We do not collect health or medical data about you personally
• We do not use the Advertising Identifier (IDFA) or engage in ad tracking
• We do not collect Biometric Data for identification or authentication purposes — it is used solely for AI avatar video and synthesized speech generation
• We do not use Presenter Assets to train general-purpose AI models owned by EagleScope

3.1 Video Generation

• Processing uploaded research papers through our classification, extraction, and script generation pipeline
• Generating AI avatar video clips from Presenter photographs via HeyGen (Avatar IV technology)
• Synthesizing voiceover audio from voice clones via ElevenLabs
• Rendering animated HTML data visualization scenes via Playwright and Replicate/Kling
• Compositing final video output via FFmpeg
• Generating multi-language video variants in up to twelve (12) languages when requested
• Generating 9:16 vertical and 16:9 landscape format variants

3.2 Service Operations

• Managing Customer accounts, orders, credits, and billing
• Delivering completed videos and processing revision requests through our surgical revision system
• Providing customer support
• Maintaining and improving the PaperVideo pipeline
• Caching Presenter Assets (photographs and voice clones) to reduce processing time and cost for subsequent video orders featuring the same Presenter

3.3 Quality and Safety

• Verifying that generated content accurately represents the source research paper
• Monitoring for fabrication or misrepresentation in AI-generated outputs
• Debugging pipeline errors using intermediate JSON files saved at each processing stage

3.4 Communications

We may use your contact information to send order confirmations, delivery notifications, script review requests, revision updates, and essential service-related communications. Marketing communications require your separate opt-in consent.

We do not sell your personal information. We do not share your personal information with advertisers.

4.1 Third-Party AI Providers

To generate PaperVideo videos, we transmit certain data to the following providers:

• HeyGen — receives Presenter photographs to generate lip-synced avatar video clips using Avatar IV technology. HeyGen processes facial geometry to animate the avatar.
• ElevenLabs — receives voice recordings to create and maintain a voice clone. ElevenLabs processes voiceprint data to synthesize per-scene voiceover audio.
• Anthropic (Claude API) — receives text extracted from research paper PDFs for classification, content extraction, and script generation. No Presenter photographs, voice recordings, or Biometric Data are sent to Anthropic.
• Kling / Replicate — receives rendered scene frames for video processing. No Biometric Data is sent to these providers.

Each Third-Party AI Provider is contractually obligated to use the data solely for the purpose of performing the requested processing and to maintain appropriate security measures. Once data is transmitted to a Third-Party AI Provider, it is also subject to that provider's own privacy policy and data practices.

4.2 Distribution of Generated Content

Completed PaperVideo videos may be distributed by the Customer through:

• Social media platforms (Instagram Reels, TikTok, YouTube Shorts, LinkedIn, X/Twitter)
• Academic conferences and grand rounds presentations (landscape 16:9 variant)
• Journal websites and publisher platforms
• Author landing pages hosted by EagleStudio
• Institutional websites and email newsletters
• Medical education and continuing education platforms

The scope of permitted distribution is governed by the Presenter Consent and Likeness Release.

4.3 Service Providers

We use additional service providers including Supabase (database hosting), Stripe (payment processing), and cloud infrastructure providers (AWS S3 for asset storage, EC2 for backend processing).

4.4 Legal Requirements

We may disclose your information as required by law, court order, or governmental request, or as necessary to enforce our Terms of Service, protect our rights, or prevent fraud.

4.5 Business Transfers

In the event of a merger, acquisition, or sale of assets, your information (including Presenter Assets) may be transferred. We will notify affected individuals before their information becomes subject to a different privacy policy.

Certain jurisdictions impose specific requirements on the collection and processing of biometric data, including Illinois (Biometric Information Privacy Act, 740 ILCS 14), Texas, Washington, and the European Union/EEA (GDPR).

5.1 What Biometric Data We Process

• Facial geometry — derived from Presenter photographs by HeyGen to generate animated avatar video clips
• Voiceprint characteristics — derived from Presenter voice recordings by ElevenLabs to create and operate a voice clone for synthesized speech

5.2 Purpose and Legal Basis

We process Biometric Data solely for the purpose of generating AI avatar videos and synthesized voiceover audio as part of the PaperVideo Service. We do not use Biometric Data for identification, authentication, surveillance, or any purpose other than video generation. Our legal basis is the Presenter's informed, written consent provided through the Presenter Consent and Likeness Release form.

5.3 Retention and Destruction

Presenter photographs and voice recordings are retained for the duration specified in the Presenter Consent and Likeness Release, or until the Presenter withdraws consent, whichever occurs first. Cached assets are deleted within sixty (60) days of consent withdrawal or account termination, subject to each provider's data retention policies.

5.4 Security

Biometric Data is protected by encryption in transit (TLS) and at rest, role-based access controls, and automated pipeline processing that minimizes human access to raw Presenter materials. Access is limited to authorized personnel and automated pipeline processes.

5.5 No Sale or Profit

We do not sell, lease, trade, or otherwise profit from Biometric Data. Biometric Data is not shared with any party other than the Third-Party AI Providers identified in Section 4.1, solely for the purpose of video generation.

6.1 Customer Data

• Account information: retained while active; deleted within 30 days of account deletion
• Uploaded research papers (PDFs): retained for twelve (12) months after the last video order referencing that paper
• Intermediate pipeline files (extraction JSON, scripts, HTML scenes, per-scene intermediates for revision): retained for six (6) months, then purged
• Completed video files: retained for twelve (12) months after delivery

6.2 Presenter Data

• Presenter photographs: retained for the duration of the consent period or until consent is withdrawn
• Voice clone (at ElevenLabs): retained until consent is withdrawn or service is terminated, plus a 60-day wind-down period
• Avatar model (at HeyGen): retained per HeyGen's data retention policies; we request deletion upon consent withdrawal
• Completed videos already distributed prior to consent withdrawal are not recalled; no new videos will be generated after withdrawal

6.3 Anonymized Data

Anonymized, aggregated data that cannot be linked back to any individual (such as total videos produced, processing times, language distribution) may be retained indefinitely.

7.1 All Users

• Access — request a copy of the personal information we hold about you
• Correction — request correction of inaccurate information
• Deletion — request deletion of your personal information, subject to legal retention requirements
• Data portability — request your data in a portable, machine-readable format
• Opt-out of marketing — unsubscribe from marketing communications at any time

7.2 Presenters — Additional Rights

• Consent withdrawal — withdraw your Presenter Consent at any time by written notice to legal@eaglescope.ai
• Access to Biometric Data records — request a written accounting of the specific Biometric Data we hold and the Third-Party AI Providers to which it has been disclosed
• Restriction of processing — request that we limit processing of your Presenter Assets to specific use cases (e.g., a single language or single paper)

7.3 California Residents (CCPA/CPRA)

California residents have the right to know, delete, correct, and opt out of sale (we do not sell personal information). To exercise these rights, contact legal@eaglescope.ai.

7.4 European Residents (GDPR)

Individuals in the EEA, UK, or Switzerland have the right to access, rectification, erasure, restriction, portability, and to object to processing. Our legal bases are: consent (Presenter Assets and Biometric Data), contract performance (Customer orders), legitimate interests (service improvement and security), and legal obligations. You have the right to lodge a complaint with your local supervisory authority.

7.5 Illinois Residents (BIPA)

Illinois residents have specific rights under the Biometric Information Privacy Act (740 ILCS 14), including the right to receive notice before collection of Biometric Data, provide informed written consent, request information about our retention and destruction policies, and bring a private right of action for violations. Our Presenter Consent and Likeness Release form includes the required BIPA disclosures.

We implement reasonable and appropriate technical and organizational measures to protect your information, including:

• Encryption of all data in transit (TLS/SSL) and at rest
• Secure authentication for all pipeline components and API integrations
• Role-based access controls limiting personnel access to Presenter Assets and Biometric Data
• Automated pipeline processing that minimizes human access to raw Presenter materials
• Regular security reviews and vulnerability assessments
• Intermediate pipeline files stored in access-controlled object storage with audit logging

No method of electronic transmission or storage is completely secure. While we strive to protect your information, we cannot guarantee absolute security.

Your information, including Presenter Assets and Biometric Data, may be transferred to, stored, and processed in the United States and other countries where our Third-Party AI Providers operate. By providing consent and using the Service, you consent to such transfers. Where required by applicable law (including GDPR), we ensure appropriate safeguards are in place, such as Standard Contractual Clauses or equivalent mechanisms.

PaperVideo is designed for medical professionals, researchers, and academic institutions. We do not knowingly collect personal information from individuals under the age of 18. No individual under 18 may serve as a Presenter. If you believe a minor's information has been provided to us, please contact legal@eaglescope.ai immediately.

PaperVideo videos are generated using artificial intelligence. All videos include a mandatory disclosure (in the final scene) stating that the content was AI-generated. We are committed to transparency:

• Avatar animations are AI-generated from still photographs; they are not recordings of the Presenter speaking
• Voiceovers are synthesized from voice clones; they are not recordings of the Presenter reading the script
• Data visualizations are programmatically generated from extracted paper data
• The underlying research findings presented in videos are extracted from the uploaded paper and may not capture all nuances, limitations, or context

Customers are responsible for verifying the accuracy of Generated Content before distribution and must not remove or obscure the AI-generated content disclosure.

When multi-language variants (up to twelve languages) are generated, the source paper content is translated via AI and voiceover audio is re-synthesized in the target language using the same voice clone. The translated content is processed by the same pipeline and Third-Party AI Providers described in Section 4.1. EagleScope does not guarantee the accuracy of AI translations, and Customers are responsible for reviewing translated content before distribution.

We may update this Privacy Policy from time to time. We will notify you of material changes by posting a notice within the EagleStudio app and sending an email to affected Customers and Presenters. For changes affecting Biometric Data processing, we will seek renewed consent from Presenters where required by applicable law.

If you have questions about this Privacy Policy or our data practices:

• General inquiries: support@eaglescope.ai
• Legal & privacy: legal@eaglescope.ai
• Presenter / consent withdrawal requests: include “PaperVideo Presenter” in the subject line