Privacy Policy

1. Who We Are

FrameQuery is operated by Elly Software, a company registered in the United Kingdom. For any privacy-related questions or requests, contact us at privacy@framequery.com.

Under data protection law, Elly Software is the data controller for your account data, and acts as a data processor for the video and audio content you upload for analysis.

2. Data We Collect

Account Data

• Email address
• User identifier
• Authentication credentials (managed by our authentication provider)

Uploaded Content

• Video files you upload for analysis
• Audio extracted from your videos for transcription
• Video frames extracted for visual analysis

Derived Data (Processing Results)

• Scene descriptions and captions
• Detected object labels
• Shot type and angle classifications
• Transcription text with word-level timestamps
• Anonymised speaker labels ("Speaker 1", "Speaker 2", etc.)
• Detected language

Billing Data

• Subscription tier and status
• Usage metrics (hours of video indexed)
• Payment information (processed by Polar; we do not store card details)

Technical Data

• IP addresses (retained for 30 days in infrastructure logs)
• Browser and device information

Analytics Data

• Product usage events (processed by PostHog, hosted in the EU). You can opt out of analytics at any time in your account settings.
• Ad conversion measurement (Google Ads). If you accept cookies on the website, a Google tag records that a visit from a Google ad led to a download click or a purchase. For a purchase we also send Google a one-way hash of your email address, used only to match the purchase to the ad click. The tag is not loaded if you decline, and we do not use it for personalised advertising.
• Interaction logs from the desktop app, if you accepted analytics on or after 3 September 2026. These record the app's interface layout and your clicks and navigation as data, not as video or screenshots. All text is masked and all images, video frames and thumbnails are removed on your device before anything is sent, so file names, transcripts, search text and faces are never included. Turning analytics off in Settings stops this immediately.

3. How We Use Your Data

We use the information we collect to:

• Provide, maintain, and improve our services
• Process and analyze your video and audio content
• Process billing and meter usage
• Send you technical notices and support messages
• Respond to your comments and questions
• Monitor and analyze usage patterns (with opt-out)

4. Legal Basis for Processing

We process your data under the following legal bases (GDPR Article 6):

• Contract performance (Article 6(1)(b)): Account creation, video and audio processing, billing, and all core service functionality.
• Legitimate interest (Article 6(1)(f)): Product analytics to improve the service. You can opt out at any time via your account settings.
• Consent (Article 6(1)(a)): Interaction logs in the desktop app, collected only after you accept analytics in the app and withdrawable at any time in Settings.

5. Video and Audio Content

We understand that your video and audio content may contain sensitive information. We treat this content with the utmost care:

• We only process your content to provide the services you request
• Video and audio files are deleted immediately after indexing. GCS lifecycle rules serve as a safety fallback in the unlikely event that automatic deletion fails.
• Video and audio content is processed by the service providers listed in section 8 as needed to deliver the analysis, storage and sharing features you use.
• We do not use your content for AI model training
• You maintain full ownership of all your content

Quote to Clip: Free Browser Tool

Quote to Clip opens recordings and exports selected clips in your browser. When you approve transcription, extracted audio is sent to Cloudflare Workers AI. The original video, source filenames and quote searches stay on your device. The included sample needs no upload. This tool does not require an account and does not add your files or results to a saved library.

Uploaded inputs are held in private temporary storage and deletion is requested after processing, failure or cancellation. Result access expires one hour after a job starts, and the browser requests removal once it receives the result. Failed deletions are retried; a one-day storage expiry rule is a final fallback. Storage expiry is asynchronous and is not a guarantee of physical deletion at an exact hour. Clearing the tool or closing the page removes the browser's working result; files you download remain on your device.

Cloudflare Turnstile verifies requests to help prevent abuse. The tools service uses a daily rotating, keyed hash of your IP address for usage limits rather than storing the address in its job database. Quota hashes are retained for up to two days; temporary job access records for up to two days. Platform security and operational metadata may also be processed by Cloudflare under its terms.

Session replay and automatic interaction capture are disabled for tool workspaces. If you accept website analytics, we record the tool type and whether an operation started, completed, failed or produced a download. These events exclude filenames, media, search phrases and result text. We do not use tool content to train models. See Cloudflare's Workers AI data usage policy.

6. Derived Data (Metadata)

We store metadata generated from your video and audio content (scene descriptions, object labels, transcriptions, speaker labels) to provide the search and analysis functionality of the service. This metadata:

• Is never used for our own AI training
• Is never sold
• Is processed by the service providers listed in section 8 as needed to deliver the service
• Is available to you and recipients you choose through the sharing features
• Serves as a cloud backup for the desktop app

7. Speaker Diarization

Our transcription service identifies different speakers in your audio and labels them with anonymous identifiers ("Speaker 1", "Speaker 2", etc.). This process:

• Does not create or store voice profiles or biometric templates
• Does not attempt to identify who the speakers are
• Produces anonymous numbered labels only — no identity mapping is performed or stored

Speaker diarization is not considered biometric processing under GDPR Article 9 because no biometric data is generated, stored, or used for identification purposes.

8. Data Processors

We use the following third-party services to provide FrameQuery. Each has a Data Processing Agreement (DPA) in place:

An up-to-date list is also available programmatically via our API at GET /v1/privacy/data-processors.

We will notify you at least 30 days before adding or replacing a sub-processor by updating this list.

9. International Data Transfers

Our primary processing infrastructure is hosted in the EU (Google Cloud, europe-west2 London and europe-west4 Netherlands).

Speech-to-text transcription is performed on US-based GPU instances due to GPU availability constraints. These transfers are covered by:

• Google Cloud's Standard Contractual Clauses (SCCs) and Data Processing Addendum
• The EU-US Data Privacy Framework (Google LLC is a certified participant)

Audio data sent to US infrastructure is ephemeral — it is present on the US GPU for approximately 10 to 120 seconds and is deleted immediately after transcription. The GPU instances have no public IP address and self-terminate when idle.

A Transfer Impact Assessment documenting the necessity, safeguards, and risk assessment for this transfer is available on request.

10. Data Retention

11. Your Rights

Under GDPR and UK data protection law, you have the following rights:

• Access and portability: You can export all your data at any time through the app or via the API.
• Deletion: You can delete your account through the app. This triggers a 7-day recovery window, after which all your data is permanently removed across all systems — including job data, billing records, API keys, reviews, proxies, shared indexes, analytics data, and object storage files.
• Analytics opt-out: You can disable product analytics, including interaction logs, at any time in your account settings.
• Rectification: To update your personal data, contact us at privacy@framequery.com.
• Lodge a complaint: You have the right to lodge a complaint with the Information Commissioner's Office (ICO) in the UK or the relevant supervisory authority in your EU member state.

12. Security

We implement appropriate technical and organizational measures to protect your data:

• Connections to production services use HTTPS; cloud storage uses provider-managed encryption at rest
• Stored face and voice templates and face thumbnails are encrypted on your device. Ordinary local library data and the derived biometric matching index are not encrypted by FrameQuery. See our security page for the distinction between local and cloud data protection.
• Least-privilege access controls on all services and infrastructure
• GPU processing instances are ephemeral and have no public IP
• API keys are stored as SHA-256 hashes, never in plaintext
• Rate limiting per user and per IP address
• Sensitive tokens and query parameters are redacted from all logs
• SSRF protections on all external URL fetches

13. Data Breach Notification

In the event of a personal data breach affecting your data, we will notify you without undue delay and in any event within 72 hours of becoming aware of the breach. The notification will include the nature of the breach, its likely consequences, and the measures we have taken or propose to take.

14. Changes to This Policy

We may update this Privacy Policy from time to time. We will notify you of any material changes by posting the new Privacy Policy on this page and updating the 'Last updated' date. For material changes, we will also notify you by email.

15. Contact Us

If you have any questions about this Privacy Policy, please contact us at:

privacy@framequery.com