ContextScribe
Privacy Policy
We built ContextScribe so that research stays with the person who did it. Your captured pages, notes, memories and reports belong to you. This policy sets out what the Service handles, why it needs each piece of data, where that data goes, and the choices you control. Where a section describes an optional feature, nothing in it happens unless you turn the feature on.
1. Who is responsible
ContextScribe is operated by the ContextScribe team, referred to here as the operator. Contact: rafirahman3648@gmail.com We answer support mail within three business days.
2. What we handle
Account. Your email address, an optional display name, and a password, which we store only as a salted scrypt hash. Accounts are email and password only; there is no sign-in through another service.
Session. Session credentials are stored only as SHA-256 hashes with an expiry time. The usable token lives in your browser, as an httpOnly cookie and, when you sign in from the extension, in local storage for the extension to present. We do not keep it in plaintext on our servers.
Your content. Everything you create or capture: projects, sources (a page address, its domain and the extracted text), pasted text, uploaded files, the memories and summaries the assistant derives from them, notes, decisions, tasks, reminders, chat messages, research queries and reports, and a record of activity in your workspace.
AI provider settings. If you connect your own compatible provider, we store its endpoint, model name and API key. The key is encrypted with AES-256-GCM, is shown only as a masked hint afterwards, and is used solely to send your requests to that provider.
Billing. Payments are handled by Paddle, which acts as merchant of record. We keep only the subscription state Paddle returns to us, such as plan, status and reference identifiers. Card details never reach our servers.
Operational logs. Standard request logs, meaning time, address, route and outcome, kept to secure the Service, apply rate limits and investigate abuse.
3. How the Chrome extension handles your data
The extension exists for one job: to move what you are reading into your own workspace. It does that in four distinct ways, each with a different scope, and each started by you.
3.1 Capturing a page
When you choose “Capture this page” in the side panel, or the matching item in the page’s right-click menu, the extension reads that one tab: its address, its title and the text visible on it. That material is sent to your workspace over HTTPS and appears in the capture composer, where you review and run it before anything is saved as a source. Nothing is written to your workspace by this step on its own.
3.2 Capturing the list of open tabs
When you choose “Capture ALL open tabs” from the toolbar icon’s menu, the extension reads the addresses and titles of your open tabs, up to thirty of them. It does not read the content of those tabs at that point. Your workspace fetches the text of a page only for the entries you then confirm.
3.3 Automatic capture, which is optional
Automatic capture is switched off when you install the extension and stays off until you turn it on in the popup behind the toolbar icon. It is the feature that collects the most, so it is worth being direct about what enabling it means.
Once enabled, for each page you visit the extension records the address, the title and the rendered visible text of that page, up to fifty thousand characters, and sends it to your ContextScribe workspace to be turned into saved notes. It does this whether or not the side panel is open, which is why the webNavigation, tabs and alarms permissions and the broad host access in the manifest are requested. Pages are queued only in chrome.storage.session, which Chrome clears when the browser closes, so nothing waits on disk. The popup states the above before you agree and records the moment you do.
Switching the feature off takes effect immediately and discards anything still queued but unsent, so a page visited after you turn it off cannot arrive later. Undelivered entries are never held beyond the current browser session either way.
3.4 Browser control, which is optional
If you ask the assistant to carry out a task in the browser, it can read and act inside the tab you point it at, while showing a visible indicator on the page. It pauses to ask before consequential actions such as submitting a form or uploading a file. This runs only for a task you started; the extension does not act on pages on its own.
3.5 What we do not collect
In no mode does the extension read what you type, your passwords, your cookies or another site's session tokens, and it does not take screenshots or read downloaded files. It does not capture browser-internal pages, addresses on your local network, or your own ContextScribe workspace page. It keeps no record of pages you visited that were not captured.
4. Limited Use commitment
Our use of information received from Google APIs, and of any data collected through the Chrome extension, follows the Chrome Web Store Developer Program Policies, including the Limited Use requirements. Specifically:
- We collect and use data only to provide or improve the single purpose stated in the extension's listing, which is turning the pages you read into your own cited, searchable research knowledge base, together with related operational purposes such as maintaining, securing and measuring that feature.
- We transfer data to third parties only where it is necessary to provide that purpose, to comply with the law, or to protect against malware, spam, phishing or other abuse.
- We do not use or transfer data for personalised advertising, do not sell or transfer it to advertising platforms, data brokers or information resellers, and do not use it to determine creditworthiness or for lending purposes.
- We do not allow a person to read your captured content, except with your consent to help you regain access, where the data has been aggregated and anonymised for internal operations, where it is necessary for security, or where the law requires it.
5. Changes to these practices after installation
If the extension begins to collect a category of data beyond what is described above, we disclose that change inside the extension the next time you open it, and in this policy, before the new collection takes effect. We will not treat a silent update as consent. Automatic capture was introduced on that basis and remains opt-in.
6. AI processing and recipients
- To produce memories, summaries, answers with citations and research reports, your captured content is processed by a large language model. By default this is the operator-configured system AI provider; if you connect your own provider in Settings, processing goes directly to that provider using your own API key.
- Content is sent to the AI provider only to fulfill your request (capture analysis, chat, research). It is not used to train models, is not shared with advertisers, and is never sold.
- Complete list of recipients: (a) the hosting provider that runs the Service, (b) the AI provider that processes your requests — the operator-configured platform backend or your own configured provider, (c) Paddle for billing. No other third parties receive your data. The Service embeds no advertising or analytics trackers.
7. Purposes and legal bases
- Providing the Service (contract): storing and processing your account, content, AI synthesis, exports.
- Billing (contract): subscription management with Paddle.
- Security and abuse prevention (legitimate interest): hashed passwords and sessions, rate limiting, logs.
8. Storage, retention, deletion
- Your data is stored in the operator’s production database. In transit everything is served over HTTPS. Passwords are scrypt-hashed, session tokens SHA-256-hashed, and provider API keys AES-256-GCM-encrypted at rest.
- We keep your data as long as your account exists. You can delete any individual item (project, source, memory, note, task, reminder, chat, research run) directly in the app at any time.
- Full account deletion: Settings → Account → Delete account permanently removes the account and every owned row (projects, sources, memories and their version history, decisions, tasks, chats, research, activity, providers, subscriptions, sessions) in one cascade. Deletion requires typing your account email as confirmation.
- Portability before leaving: Settings → Data exports your entire workspace (JSON backup, Markdown and PDF reports) so you can take everything with you.
9. Your rights
Depending on your jurisdiction you may have rights to access, correct, export or erase your personal data, and to object to or restrict certain processing. Most of these are self-service in the app (edit memories, delete items, export backups, delete account); for anything else, contact us and we will act on valid requests within applicable legal timelines.
10. Sharing with other users
Nothing is shared with anyone by default. If you explicitly generate a project share code and hand it to another user, they can read a snapshot of that project and import a copy into their own workspace. Share codes can be revoked at any time and revoked codes stop working immediately.
11. Children
The Service is not directed to children under 13 (or the equivalent minimum age in your jurisdiction) and we do not knowingly collect their personal data. If you believe a child uses the Service, contact us so we can remove the account.
12. International transfers
Your data is processed by our hosting and AI providers, which may operate in countries other than yours. Where required, transfers rely on appropriate safeguards such as standard contractual clauses.
13. Changes to this policy
If we change this policy we will update the “Effective date” above and, for material changes, notify you inside the app before they take effect.
14. Contact
Privacy questions, data requests or complaints: If we cannot resolve your concern, you may escalate to your local data protection authority.