Introduction
LangCapture helps software professionals understand and produce English for real work, then turn selected expressions into active learning. This policy explains what we collect, what is processed only for the current request, what you choose to save, and which service providers receive data.
Information We Collect
- Account information: name, email address, profile image, authentication-provider identifiers, language preferences, time zone, reminder preferences, and account or subscription state.
- Content you deliberately submit: selected text, a draft, a communication intent, or optional context that you explicitly include in a Work or immersive-translation request.
- Content you explicitly save: saved work expressions, translations, communication intent, transferable pattern, tone guidance, learning targets, Replay attempts and assessment evidence, and TTS audio requested for reusable content.
- Usage and device data: IP address, browser type, device and session identifiers, pages viewed, product actions, error metadata, and timestamps. Product analytics are designed not to contain submitted source text, translations, drafts, or generated expressions.
- Billing records: Creem handles payment details. We retain provider customer and order identifiers, subscription status, amounts, currency, and transaction status needed to operate billing.
- Support and validation research: messages or attachments you send to support, plus written feedback or interview notes you voluntarily provide. Validation participation is stored separately from routine product personalization.
We do not ask the product to collect your job title, employer, claimed frequent work scenarios, or company profile for personalization. We do not automatically add your username or account profile to model prompts.
Request Processing and Saving
Selected content, optional context, drafts, full-page translation text, and unaccepted model candidates are processed for the current request but are not added to a server-side task history. A long-term personal asset is created only when you choose Save Result or Make It Yours.
An unsaved TTS result is temporary and is scheduled for deletion after 24 hours. Audio for an expression you deliberately saved may be retained for replay until the expression, audio, or account is deleted. Immersive page translations remain in the current webpage and are not saved as page archives.
AI and TTS Processing
Content-bearing language-model requests are routed through OpenRouter. LangCapture requests denial of provider data collection and requires Zero Data Retention for eligible routing endpoints. The request fails if those privacy conditions cannot be satisfied; the product does not silently downgrade them. The selected downstream model provider receives the submitted request through OpenRouter under those routing requirements.
MiniMax processes only the text you explicitly ask LangCapture to speak. We do not append your communication intent, optional work context, account profile, or unrelated page content to a TTS request.
You must not submit credentials, secrets, personal information, confidential company information, or other content you are not authorized to process through third-party AI services.
How We Use Information
We use information to provide authentication, Work assistance, personal-expression matching, Replay and learning scheduling, TTS, immersive translation, data export and deletion, billing, reminders you enable, support, security, abuse prevention, reliability, and privacy-safe product analytics.
We do not use submitted work content to create advertising profiles. Under the model-routing configuration described above, content is not submitted for provider training or retained by eligible model endpoints. We do not claim that automated redaction makes confidential content safe to submit.
Service Providers
- Cloudflare: Worker hosting, D1 database, R2 audio storage, and related security and delivery infrastructure.
- OpenRouter and privacy-qualified downstream model providers: language-model inference under the privacy gate described above.
- MiniMax: text-to-speech synthesis for text you explicitly request to hear.
- Creem: checkout, payment processing, subscriptions, and customer billing portal.
- Resend: verification, security, learning-reminder, and other transactional email.
- Google Analytics: privacy-filtered acquisition and product events when configured; work content is excluded.
- Google and GitHub: social sign-in when enabled and chosen by you.
We may also disclose necessary information to comply with law, enforce our terms, investigate abuse, or protect users and systems.
Retention and Control
We retain account data and explicitly saved personal assets while your account is active or as needed to provide the service. When a trial or subscription ends, your personal library becomes read-only but remains viewable, exportable, and deletable. You may delete individual personal expressions, export your library, or request account deletion. Limited billing, fraud-prevention, legal, and backup records may remain where required.
Security and International Transfers
We use reasonable technical and organizational safeguards, including authenticated access, scoped extension tokens, provider privacy routing, limited content-bearing analytics, and operational abuse controls. No transmission or storage system is completely secure. Our providers may process data outside your country; we use appropriate safeguards where required by law.
Your Choices and Rights
Depending on your location, you may have rights to access, correct, export, restrict, or delete personal data. You can disable Replay reminders and weekly summaries independently, manage cookies in your browser, cancel a subscription, and request assistance at support@langcapture.com.
Changes and Contact
We may update this policy as the product or providers change. We will publish the revised policy and update the effective date above. Questions or privacy requests can be sent to support@langcapture.com.