Version 2026-05-30

Privacy Policy

Effective 2026-05-30. Last updated 2026-05-30.

oThought is a private, call-first AI thought partner. Calls, transcripts, ideas, plans, research findings, usage, and billing state can all be personal data.

We use voice, AI, hosting, billing, app-store, notification, support, and security providers to run the service.

Users can delete sensitive content separately from deleting the full account. Full account deletion is available from Account settings.

Who this policy is for

This Privacy Policy explains how oThought collects, uses, shares, and protects personal data when someone uses the oThought mobile app, website, support channels, and related services.

oThought is a voice-first AI thought partner for private thinking, idea capture, planning, research, and follow-up.

oThought is operated by Taskless Eenmanszaak, operating from Netherlands. The current operator address is Box C9857 Verdunplein 17, 5627SZ, Eindhoven, Netherlands.

Operator registration: KVK 97491071; VAT NL005273102B84.

Personal data we collect

We collect the data needed to run a call-first AI companion and keep the user's idea system continuous across calls and devices.

  • Account data, including user ID, sign-in provider, email address, phone number if provided, display name, authentication metadata, and account status.
  • Voice and call data, including microphone permission state, live-call audio transport, recordings when used, transcript chunks, timestamps, speaker labels, provider status, and call end reasons.
  • Thought and idea data, including saved ideas, source quotes, transcripts, summaries, plans, tasks, research findings, user-stated facts, AI inferences, confidence labels, graph links, revisions, and follow-up context.
  • Usage and billing data, including plan, entitlement, voice-minute usage, top-ups, purchase status, usage ledger rows, estimated provider cost, and store or RevenueCat transaction identifiers.
  • Device and app data, including anti-abuse identifiers, app version, build variant, crash or diagnostic events, connectivity status, notification tokens, and basic logs.
  • Support and legal data, including support messages, deletion requests, privacy-rights requests, policy notices, and operational records needed to answer those requests.

Microphone, audio, and transcripts

oThought asks for microphone access so the user can talk to the AI companion.

When a call or voice capture runs, audio may be streamed to server-side voice infrastructure and third-party AI or speech providers so they can transcribe, respond, detect ideas, and produce speech.

Transcripts and selected context may be saved so ideas can be recovered after the call and used in later briefings.

How we use personal data

We use personal data to provide, secure, improve, and support oThought.

  • Start and maintain live AI calls, including ringing, connected, transcript, interruption, and post-call states.
  • Detect, save, update, merge, and display ideas in the saved idea list and Mastermind Map.
  • Generate background summaries, plans, follow-ups, research findings, and next-call briefings.
  • Sync account data across devices and restore the user's idea history after sign-in.
  • Track usage, voice minutes, cost, plan entitlements, top-ups, fraud prevention, and purchase support.
  • Send user-requested notifications, important account notices, legal update notices, and service messages.
  • Maintain security, debug failures, prevent abuse, respond to support requests, comply with law, and enforce the Terms.

AI processing and model limits

oThought uses AI systems to transcribe, respond, detect ideas, summarize, infer relationships, generate plans, and research public sources. AI output can be incomplete, wrong, or out of date.

We do not sell private transcripts or ideas. We do not use private transcripts, recordings, or ideas for public examples or advertising without separate permission.

Before public launch, every AI provider's data-retention and training settings should be verified and configured to avoid training on user content where provider controls allow.

Legal bases for processing

Where laws such as the GDPR or UK GDPR apply, we rely on different legal bases depending on the processing activity.

  • Contract: to provide calls, idea capture, sync, plans, billing, subscriptions, support, and account features.
  • Consent: for microphone access, optional notifications, optional marketing, and processing that legally requires consent.
  • Legitimate interests: to secure the service, prevent abuse, improve reliability, debug issues, measure usage, and protect contribution margin, balanced against user privacy.
  • Legal obligation: to retain records needed for tax, accounting, fraud prevention, consumer protection, legal claims, and regulatory compliance.

Sharing and processors

We share personal data only as needed to operate oThought, comply with law, or protect users and the service.

Categories of recipients may include cloud hosting, database, authentication, voice, transcription, AI, research, notification, analytics, crash reporting, payment, app store, customer support, and security providers.

The Privacy Policy should describe provider categories and link to the current provider/subprocessor inventory. The provider inventory can change more often than the Terms, but material changes to data use, training, retention, region, or sensitive-content access may require user notice or acceptance.

Current provider inventory

Current provider inventory reviewed 2026-05-27: 9 active or platform providers, 9 configurable or fallback providers.

Some providers are only used when a runtime setting, fallback key, platform, or build environment enables them.

  • Convex: backend hosting, database; data handled: account identifiers, calls and transcript rows, saved ideas, plans, tasks, graph data, and research findings.
  • LiveKit Cloud: realtime voice transport, LiveKit agent runtime; data handled: live call audio transport, session metadata, agent messages and operational logs.
  • Google Gemini API / Google AI: realtime conversation engine, compose STT/TTS; data handled: audio or transcript-derived prompts, conversation context, AI outputs.
  • xAI API: Grok realtime voice option, compose STT/TTS option; data handled: audio or transcript-derived prompts, conversation context, AI outputs.
  • GroqCloud: compose STT option, compose TTS option; data handled: recorded audio or transcript-derived prompts when selected, AI audio/text outputs, usage metadata.
  • Anthropic Claude API: background text LLM option; data handled: transcript-derived prompts, idea context, plans, research, and generated outputs.
  • Operator-configured OpenAI-compatible LLM endpoint: background text LLM option, research LLM option; data handled: transcript-derived prompts, idea context, research prompts.
  • Superwall: subscription and in-app purchase entitlement management, purchase webhooks; data handled: app user id, purchase identifiers, subscription state.
  • Apple: App Store distribution, Sign in with Apple; data handled: Apple sign-in identity or private relay email, subscriber id and purchase reports, country/region purchase metadata.
  • Google / Google Play: Google sign-in, Google Play distribution; data handled: Google sign-in identity, purchase and entitlement metadata, Play Integrity request hashes, package metadata, signing certificates, license status, and device integrity signals.
  • Expo / EAS / Expo Push Service: native builds, public build variables; data handled: push tokens, notification payloads, build metadata.
  • Cloudflare Email Service: email OTP, domain email infrastructure; data handled: email address, OTP email content, delivery metadata.
  • Resend: email OTP fallback; data handled: email address, OTP email content, delivery metadata.
  • Twilio SendGrid: email OTP fallback; data handled: email address, OTP email content, delivery metadata.
  • Telnyx: phone OTP; data handled: phone number, OTP SMS content, delivery metadata.
  • Twilio: phone OTP fallback; data handled: phone number, OTP SMS content, delivery metadata.
  • Infisical: secret management, environment sync source of truth; data handled: provider API keys, deployment configuration, operator account data.
  • Seq-compatible log sink: server log aggregation; data handled: sanitized server events, error metadata, operational diagnostics.

International transfers

oThought may process data in countries other than the user's country.

Where required, we use appropriate safeguards such as standard contractual clauses, data-processing agreements, transfer assessments, or equivalent mechanisms.

Retention

We keep user content for as long as the account is active or as needed to provide the service, unless the user deletes it or requests deletion.

Calls, transcripts, recordings, ideas, plans, tasks, research findings, and graph data should be deletable from user-visible surfaces. After account deletion, retained categories may include: billing, tax, accounting, refund, and chargeback records; fraud-prevention, abuse-prevention, and security records; support, privacy-rights, deletion-audit, legal, and dispute records.

Deletion and privacy rights

Users can delete calls, transcripts, recordings, ideas, plans, and related app content from Settings without deleting their account.

Depending on location, users may have rights to access, correct, delete, export, restrict, object to, or appeal decisions about their personal data. Privacy-rights requests should go to privacy@othought.com; the EU/UK GDPR target response window in the current operator profile is 30 days.

Deleting an oThought account does not automatically cancel an active Apple or Google subscription. Users should cancel subscriptions through the relevant store's subscription-management tools.

Children and age

oThought is not designed for children. Users must be at least 16 years old and old enough to consent to data processing in their country.

oThought is not directed to children and does not knowingly allow users under 13.

Sensitive content and safety

oThought may contain sensitive thoughts. The app is not therapy, medical care, legal advice, financial advice, or emergency support.

If the system detects crisis, self-harm, or imminent safety signals, it should respond supportively and point the user to emergency services or crisis resources.

Security

We use technical and organizational safeguards appropriate for a private AI thought product, including server-side secret storage, authenticated backend access, provider isolation, deletion controls, and limits on raw transcript logging.

No system is perfectly secure. Users should avoid sharing passwords, credentials, illegal content, or information they do not have the right to disclose.

Changes to this policy

We may update this Privacy Policy as the product, providers, laws, or business model changes.

For material changes, we should notify users in the app, and where required by law also by email, push notification, store notice, or affirmative consent before the change takes effect.

Contact

Privacy questions and requests should go to privacy@othought.com. General support should go to support@othought.com. Data protection contact: privacy@othought.com.

Current operator profile: Taskless Eenmanszaak, Box C9857 Verdunplein 17, 5627SZ, Eindhoven, Netherlands.

Operator registration: KVK 97491071; VAT NL005273102B84.