Introduction

AI Video Summarizer ("the Extension", "we", "our") is committed to protecting your privacy. This Privacy Policy explains how the Extension handles information when you use it to generate AI-powered summaries of video content.

The core principle of this policy is simple: the Extension processes the minimum data necessary to generate your summary and does not store, sell, or share any of your data beyond what is strictly required for the summarization to work.

Data We Process

The Extension processes the following data, and only when you explicitly click the "Summarize" button:

Transcript (subtitles / closed captions). The text content of subtitles or captions extracted from the video page. This is the primary input for AI summarization. For YouTube, Vimeo, and Loom, transcripts are extracted locally from the video page. For Facebook, transcript extraction is performed by our server (see "Facebook Transcript Processing" below).

Video title. The title of the video, extracted from the video page's metadata. Used to provide context to the AI model for better summary quality.

Video URL. For YouTube, Vimeo, and Loom: the URL is processed entirely on your device to identify the platform and fetch the transcript. It is not transmitted to our servers. For Facebook: the URL is sent to our Firebase Cloud Function, as Facebook video transcript extraction cannot be performed from the browser alone.

Facebook authentication cookies. When you summarize a Facebook video, the Extension reads three specific Facebook cookies from your browser—c_user (user ID), xs (session token), and datr (browser integrity token)—and transmits them to our Facebook Transcript Cloud Function. These cookies are used solely to allow our server to access the video on your behalf (Facebook requires authentication for some video content). No other cookies are accessed or transmitted.

Audio data (Facebook only). If a Facebook video has no text captions available, our Cloud Function downloads the video's audio track and sends it to OpenAI's Whisper API for speech-to-text transcription. The audio is processed in memory on the server, transcribed, and immediately deleted. No audio is retained.

OpenAI API key (optional). If you choose to provide your own OpenAI API key, it is stored in chrome.storage and used to call the OpenAI API directly from your browser, bypassing our Firebase proxy. The key is stored locally on your device and is never transmitted to our servers.

What we do NOT collect:

• Personal identity information (name, email, address, etc.)

• Browsing history

• IP addresses or location data

• Device or browser fingerprints

• Tracking identifiers or advertising IDs

• Any data from pages other than the video you explicitly submit

• Facebook cookies other than the three listed above (c_user, xs, datr)

How We Use Data

YouTube, Vimeo, and Loom videos. The transcript text and video title are sent to our Firebase Cloud Function (openaiProxy), which forwards them to the OpenAI API (GPT-4o-mini). The function is a stateless proxy—it receives the request, forwards it to OpenAI, returns the generated summary, and discards all data. No transcript, title, or summary is stored on any server, database, or log.

Facebook videos. The video URL and authentication cookies are sent to our Facebook Transcript Cloud Function. This function uses yt-dlp (an open-source tool) to extract subtitles and metadata from the video. If text captions are unavailable, the function downloads the audio and sends it to OpenAI's Whisper API for transcription. Once the transcript is obtained, it is sent (along with the video title) to our openaiProxy function, which forwards it to OpenAI's GPT-4o-mini for summarization. After the summary is returned, all data—URL, cookies, transcript, audio, summary—is immediately discarded. No copies are retained.

User-provided API key. If you have configured your own OpenAI API key, the transcript and video title are sent directly from your browser to OpenAI's API, bypassing our Firebase Cloud Functions entirely.

In all cases, once the summary is displayed to you, the transcript data is immediately discarded from memory on your device. No copies are kept on any server, database, or log.

Facebook video transcript extraction requires server-side processing because Facebook does not expose subtitle data through browser APIs. Here is exactly what happens:

1. When you initiate summarization of a Facebook video, the Extension reads your Facebook authentication cookies (c_user, xs, datr only) using the browser.cookies API.
2. The video URL and these cookies are transmitted via HTTPS to our Facebook Transcript Cloud Function (facebookTranscript).
3. The Cloud Function uses yt-dlp (bundled open-source binary) to extract video metadata (title, thumbnail, duration) and attempt subtitle download.
4. If text subtitles (SRT format) are found, they are parsed and returned.
5. If no text subtitles are available, yt-dlp downloads the audio track (compressed to 64kbps MP3), which is then sent to OpenAI's Whisper API for transcription. The audio is read into memory, transmitted to Whisper, and the resulting transcript segments are returned.
6. All temporary files (SRT files, audio files, cookie files) are deleted immediately after processing. No audio, transcript, or cookie data persists on the server after the request completes.
7. The transcript segments, title, and metadata are returned to the Extension, which then proceeds with AI summarization as described above.

The cookies are transmitted in Netscape cookie file format, base64-encoded, over HTTPS. They are written to a temporary file on the Cloud Function's filesystem (for yt-dlp compatibility), used for the duration of the request, and deleted immediately after.

Third-Party Services

The Extension relies on the following third-party services:

OpenAI API (GPT-4o-mini). The transcript text and video title are sent to OpenAI's API to generate the summary. When using our server-side proxy, the request originates from our Firebase Cloud Function. When using your own API key, the request originates from your browser. OpenAI's API usage is subject to OpenAI's Privacy Policy (https://openai.com/policies/privacy-policy). Per OpenAI's API data usage policies, data sent via the API is not used to train OpenAI's models.

OpenAI Whisper API. For Facebook videos without text captions, audio data is sent to OpenAI's Whisper API for speech-to-text transcription. This is subject to the same OpenAI Privacy Policy and data usage terms as above.

Firebase Cloud Functions (Google). Our server-side functions run on Google's Firebase platform. The openaiProxy function receives transcript text and forwards it to OpenAI. The facebookTranscript function receives the video URL and authentication cookies to extract the transcript. Both functions are stateless—they do not store any user data. Error logs (containing error types and HTTP status codes, not your content) may be retained by Firebase according to Google's logging policies. Firebase is subject to Google's Privacy Policy (https://policies.google.com/privacy).

Data Storage

Server-side. Neither of our Firebase Cloud Functions stores any user data. They are stateless proxies—they receive the request, process it, return the result, and discard all data. No database, file storage, or caching layer retains transcripts, summaries, URLs, cookies, audio, or any other user content.

Local storage (chrome.storage). The Extension stores two optional values:

• firebase_function_url — a technical configuration value that can override the default Cloud Function endpoint. Only

present if you have configured a custom deployment.

• openai_api_key — your personal OpenAI API key. Only stored if you have explicitly entered one in the Extension settings.

No transcripts, summaries, video URLs, usage history, or browsing data are stored on your device.

Analytics and Tracking

The Extension does not use tracking, or crash-reporting services. We do not use Firebase Analytics, Sentry, Crashlytics, or any similar service. We do not track how you use the Extension, which videos you summarize, or how often you use it. There are no hidden trackers, pixels, or fingerprinting scripts.

Your Control and Data Deletion

Because the Extension does not store your data, there is no data to access, export, or delete. Each summarization request is ephemeral — the transcript is processed in memory and discarded immediately after the summary is returned to you.

Children's Privacy

The Extension is not directed at children under the age of 13. We do not knowingly collect personal information from children. As noted throughout this policy, the Extension does not collect personal information from any user.

We do not sell, rent, trade, or share any user data with any third party. The data processed by the Extension is transmitted only to the services described above (Firebase Cloud Functions and OpenAI API) and solely for the purpose of generating your video summary. No data is made available to advertisers, data brokers, analytics companies, or any other entity.

The Extension requests the following browser permissions, each for a specific technical purpose:

• storage — to save your optional OpenAI API key and custom function URL configuration locally.

• scripting — to extract video transcripts and metadata from YouTube, Vimeo, and Loom video pages.

• tabs — to open the Extension's interface page when you click the Extension icon.

• cookies — to read Facebook authentication cookies (c_user, xs, datr only) for accessing Facebook video subtitles that require authentication.

Host permissions:

• api.openai.com — to send transcripts directly to OpenAI when using your own API key, and to use Whisper for Facebook audio transcription.

• youtube.com, vimeo.com, loom.com — to extract transcripts from videos on these platforms.

• facebook.com, fb.watch — to extract Facebook video transcripts, including reading authentication cookies.

• cloudfunctions.net — to communicate with our Firebase Cloud Functions.

Changes to This Privacy Policy

We may update this Privacy Policy from time to time. When we do, we will update the "Last updated" date at the top of this policy. We encourage you to review this policy periodically. Continued use of the Extension after changes constitutes acceptance of the updated policy.

Contact

If you have questions about this Privacy Policy or the Extension's data practices, please contact us at cwshacks1@gmail.com