Teddy Browser Extension Privacy Policy
Teddy is a browser extension and backend service that helps students ask natural-language questions about Panopto lectures. This Privacy Policy explains what information we collect through the extension and supporting APIs, why we collect it, how we use it, and the choices you have. By installing or using the extension you agree to this Policy.
1. Information We Collect
| Category | Details | Where it lives |
|---|---|---|
| Questions you type | Prompts sent via the widget UI, including follow-up messages. |
Stored temporarily in memory; prior chats persist in the
browser’s localStorage until you clear them.
|
| Page & playback context |
Current page URL, detected video_id, and the
timestamp you are watching so answers can cite exact moments.
|
Included in inference requests and stored with chat history locally. |
| Captions and video metadata | Transcript rows (timestamp + text), optional title, and duration scraped from the active Panopto page. | Uploads to the API, sent externally for embedding, and then stored in our database. |
| Usage metrics | Query counts, caption uploads, and aggregated statistics (no content) recorded via our backend. | Stored in our database for dashboards. |
The extension does not access your browser history, passwords, or non-Panopto pages. We do not knowingly collect special-category data unless it appears inside lecture captions or your own prompts. The extension servers do NOT store conversation history. We do not tie individual messages or conversations to individual users.
2. How We Use Information
- Answer your questions. Our backend sends your prompt, relevant transcript snippets, and optional playback timestamp to our AI provider for generation.
- Prepare each lecture. Captions are chunked, embedded via our embedding provider, and stored so future questions can be answered quickly.
- Maintain and improve Teddy. Aggregated metrics and logs help monitor uptime, tune retrieval quality, and defend against abuse.
We do not build advertising profiles, sell personal information, or use your data for marketing.
3. Local Storage & Controls
-
Conversation history is saved per domain/video inside
your browser's
localStorage. Use the “Clear conversation” button to remove it. - A flag prevents duplicate caption uploads; delete it via your browser's storage inspector if needed.
4. Data Sharing & Processors
We share information only with service providers needed to operate Teddy:
- OpenAI and other inerence providers receive your prompt, limited chat history, and retrieved transcript excerpts to generate answers.
- Voyage AI receives prompt and transcript chunks during ingestion to create embeddings.
- Hosting & infrastructure partners (e.g., Fly.io, PostgreSQL, storage, logging) store encrypted copies of databases and logs.
These vendors act as processors under contractual confidentiality and security commitments. We do not permit them to use Teddy data for their own advertising or model training beyond what their base terms require.
5. Data Retention
- Captions, embeddings, and summaries remain in our database indefinitely. There is no automatic expiration because they are needed to answer future questions.
- Usage metrics are aggregated and retained for trend analysis.
- Local chat history is retained only in your browser.
6. Changes to This Policy
We may update this Policy as the project evolves. We will refresh the “Last updated” date above and, when feasible, announce material changes. Continued use after changes become effective constitutes acceptance.