Privacy Policy

This Privacy Policy explains how VideoInsights (“VideoInsights,” “we,” “us,” or “our”), operating at videoinsights.ai, collects, uses, stores and shares personal data when you use:

• our public website at www.videoinsights.ai;
• the VideoInsights web application at chat.videoinsights.ai;
• the VideoInsights REST API and SDKs;
• the VideoInsights app for ChatGPT and Codex, which connects your VideoInsights account to OpenAI products through the Model Context Protocol (MCP) at mcp.videoinsights.ai;

(together, the “services”).

VideoInsights is the data controller for the personal data described in this policy. We are based in British Columbia, Canada. You can reach us at contact@videoinsights.ai or through www.videoinsights.ai/contact.

Our use of cookies on the public website is described separately in our Cookie Policy.

Table of Contents

• 1. Data we collect
• 2. How we use your data
• 3. AI processing
• 4. ChatGPT and Codex app
• 5. Service providers and data sharing
• 6. Data retention
• 7. Your rights and choices
• 8. Security
• 9. International transfers
• 10. Children
• 11. Changes to this policy
• 12. Contact us

1. Data we collect

Account data

When you create an account we collect your email address and, if you sign in with a social identity provider such as Google, the name and profile identifiers that provider shares with us. Authentication is handled by Amazon Cognito (see section 5). We assign your account internal identifiers and record your plan and subscription state.

Content you provide

• Media files you upload (video and audio), together with their file name, type, size and duration.
• YouTube videos you ask us to work with, identified by their video ID or URL, and the list of videos saved in your YouTube library.
• Transcripts, subtitles and translations we generate from your media or from YouTube videos.
• Questions, prompts and chat messages you send about a video, and the AI outputs (answers, summaries, metadata, visual analysis) we generate in response.
• Feedback and support messages you send us through the app, the contact form or email.

You retain full ownership of the content you upload. We do not claim any rights over your files or over the outputs we generate for you, and we never use your content to train AI models.

Billing data

Payments are processed by Stripe. We receive and store your Stripe customer and subscription identifiers, your plan, billing status and invoice events. We never see or store your full card number.

Usage and technical data

For every request to our API, web app or MCP server we record the account, the operation performed (for example “transcript”, “summary”, “ask”), the channel it came from (web app, API key, ChatGPT/Codex), a timestamp, processing times, job status and the number of VI tokens consumed. Our servers also keep standard access logs including IP address, user agent and error details.

On the public website we use Google Analytics to understand aggregate visitor behaviour. See the Cookie Policy for details and opt-out options.

2. How we use your data

We use the data above to:

• provide the services: transcribe, analyse and answer questions about your media and about YouTube videos;
• authenticate you and keep your account secure;
• meter VI-token usage, enforce plan limits and bill you;
• respond to support requests and feedback;
• monitor reliability, debug failures and prevent abuse;
• send service notices, such as changes to these terms or to your plan; and
• comply with legal obligations.

We do not sell personal data, we do not use it for advertising, and we do not use your content or your conversations to train machine-learning models.

3. AI processing

Transcription and language-model inference run on infrastructure that we operate ourselves: speech-to-text uses an open-source Whisper model and text and visual analysis uses open-weight language and vision models (currently NVIDIA Nemotron and Qwen families) served from our own hardware. Your media, transcripts and prompts are not sent to third-party AI model providers to generate results.

For YouTube videos, we first try to obtain the publisher's existing captions through the YouTube Data API or a captions provider (see section 5); only the public video ID is shared for that purpose. If no captions are available we download the audio and transcribe it ourselves.

To debug quality problems we keep traces of model calls (prompts, transcript excerpts and model outputs) in a self-hosted tracing tool on our own infrastructure. Access is restricted to our engineering team and these traces are subject to the same retention rules as the underlying content.

4. ChatGPT and Codex app

The VideoInsights app lets you use VideoInsights from inside OpenAI products (ChatGPT and Codex). It connects to our MCP server at mcp.videoinsights.ai. OpenAI is a separate controller of the data you enter into ChatGPT or Codex; their handling of your conversations is governed by OpenAI's privacy policy.

Account linking

When you connect the app you sign in to VideoInsights and approve a consent screen listing the permissions requested. The app can ask for two scopes: vi.read (see your media library, transcripts and account balance) and vi.analyze (analyse videos and media for you, using VI tokens from your plan). We then issue OAuth tokens to OpenAI. Access tokens expire after one hour and refresh tokens after 30 days of inactivity. We store only cryptographic hashes of these tokens.

What OpenAI receives

When ChatGPT or Codex calls one of our tools on your behalf, the result is returned to OpenAI and becomes part of your conversation there. Depending on the tool, results may contain:

• Profile: an opaque VideoInsights account identifier and the email address on your account.
• Account: your plan name, VI-token allowance, usage and remaining balance, the reset date and your plan's upload limits.
• Media library: the file names, types, sizes, durations, upload dates, scheduled deletion dates and transcription status of your uploaded media.
• Transcripts and analysis: transcripts of your media and of YouTube videos, and the answers and summaries our models generate about them.
• YouTube data: public video metadata (title, channel, description, statistics) and the text of public comments. We strip commenter names before returning comments.
• Job status: identifiers and status of long-running analysis jobs.

Tool results never include internal database identifiers, request or trace identifiers, or other users' data.

What we receive from OpenAI

We receive the tool name and arguments ChatGPT or Codex sends (for example a YouTube URL, a media ID or your question), together with your OAuth token. We do not receive the rest of your conversation. These requests are logged as described in section 1 and tagged as coming from the ChatGPT/Codex channel.

Disconnecting

You can disconnect VideoInsights at any time from the apps or connectors settings in ChatGPT or Codex. You can also email contact@videoinsights.ai and we will revoke all tokens issued to OpenAI for your account. Deleting your VideoInsights account revokes them as well. See our support page for step-by-step instructions.

5. Service providers and data sharing

We share personal data only with the providers below, who process it on our behalf and under contractual confidentiality obligations, or where the law requires us to. We do not sell or rent personal data.

YouTube content. When you work with a YouTube video, we retrieve public data about that video from YouTube. Your use of these features is also subject to the YouTube Terms of Service and the Google Privacy Policy. Comments we return are those of third parties; we remove commenter names and do not store comments beyond what is needed to answer your request.

Legal requirements. We may disclose personal data if required by law, regulation, legal process or a governmental request, or to protect the rights, property or safety of VideoInsights, our users or the public. If VideoInsights is involved in a merger, acquisition or asset sale, your data may be transferred as part of that transaction and we will notify you before it becomes subject to a different policy.

6. Data retention

Uploaded media files

Uploaded media files are automatically deleted after a retention period that depends on your plan at the time of upload. Each file's scheduled deletion date is shown in your library.

You can delete any file earlier from your library or through the API. When a file is deleted, its transcript, summaries, visual analysis and other derived outputs are deleted with it.

Other data

• YouTube transcripts and metadata are cached so that repeat requests for the same public video are fast and do not consume extra tokens. They are not tied to your media retention period. Videos you remove from your YouTube library are no longer associated with your account.
• Chat history in the web app is kept until you delete the conversation or your account.
• Account, plan and usage records are kept for as long as your account exists.
• OAuth tokens for the ChatGPT and Codex app expire as described in section 4 and are removed when you disconnect or delete your account.
• Billing records are retained for as long as tax and accounting law requires, which may extend beyond account deletion.
• Server logs and model traces are rotated on a rolling basis and are not kept longer than needed for debugging and security.

When you delete your account we delete your media, transcripts, chat history, tokens and account record, and we retain only what the law requires us to keep.

7. Your rights and choices

Depending on where you live, you may have the right under laws such as Canada's PIPEDA, the EU and UK GDPR or US state privacy laws to access, correct, export or delete your personal data, to object to or restrict certain processing, and to lodge a complaint with a supervisory authority. We honour these requests for all users regardless of location.

• Delete media or chats at any time from the web app or through the API.
• Disconnect the ChatGPT or Codex app from the connector settings in those products, or ask us to revoke access.
• Revoke API keys from your account settings.
• Delete your account from the security settings in the web app, or by emailing us.
• Access or export your data, or ask a question about this policy, by emailing contact@videoinsights.ai. We respond within 30 days.
• Opt out of analytics as described in the Cookie Policy.

We may need to verify your identity before acting on a request. Our legal bases for processing, where the GDPR applies, are performance of our contract with you, our legitimate interests in operating and securing the services, compliance with legal obligations and, for optional analytics cookies, your consent.

8. Security

Data is encrypted in transit (TLS) and at rest. Media files are stored in private storage buckets and served through short-lived signed URLs. Passwords are handled by Amazon Cognito and are never visible to us. OAuth tokens are stored only as hashes. Access to production systems is restricted to authorised team members. No system is perfectly secure; if we learn of a breach affecting your data we will notify you and the relevant authorities as required by law.

9. International transfers

We are located in Canada. Our own processing infrastructure is in Canada, and our service providers listed in section 5 operate in Canada, the United States and the European Union. Where personal data leaves your jurisdiction we rely on the providers' standard contractual safeguards.

10. Children

The services are not directed to children, and you must be old enough to form a binding contract in your jurisdiction to use them (see our Terms & Conditions). We do not knowingly collect personal data from children under 16. If you believe a child has provided us with personal data, contact us and we will delete it.

11. Changes to this policy

We may update this policy to reflect changes to the services or to the law. We will post the new version here and update the “last updated” date. For material changes we will also notify you by email or in the web app.

12. Contact us

VideoInsights, British Columbia, Canada

www.videoinsights.ai/contact
contact@videoinsights.ai