Privacy Policy

1. Data Collection and Use

Spokenly operates with a privacy-first approach. If you purchase a subscription through our website (via Polar or Paddle), you will need to sign in with the email address you used at checkout to activate your subscription. We use your email for authentication, subscription management, emails about your subscription and occasional emails about Spokenly Pro. Every such email has an unsubscribe link.

Our in-house analytics:

• Only collects basic app usage data (button clicks and page views inside app)
• Never includes any audio recordings or transcribed content
• Is used to improve app functionality and user experience
• Is linked to your account only when you are signed in

2. Audio and Transcription

Our analytics tracking does not collect or analyze any details about the content of your spoken audio. Your speech content remains private and is only processed for transcription purposes.

When using local models: Your audio recordings and transcriptions never leave your device. All speech-to-text processing happens entirely on your device, ensuring complete privacy of your spoken content. Note: When a Mode uses AI Instructions, your transcription and any optional context for that Mode may be sent to AI services for text processing, even if transcription itself runs locally.

When using built-in cloud models (no API key required): Spokenly captures your voice input and sends the corresponding audio file to our backend for transcription purposes. Our backend does not store these recordings. Instead, it immediately forwards them to third-party transcription services to perform the speech-to-text transcription.

When using a model with your own API key: The app sends your audio and text directly from your device to the provider you selected. This data never passes through Spokenly's servers. Spokenly's servers are only involved for built-in features that work without an API key.

When using the Spokenly plugin for ChatGPT: The plugin does not require a Spokenly account. It processes the recordings you upload in the plugin or attach in ChatGPT for transcription:

• Data: The audio, its file details (name, ChatGPT file ID for attached files, and duration), and the resulting transcript with timestamps. Uploads and playback requests also send your IP address to our servers.
• Purposes: Transcribing the recording, playing it back next to the transcript, letting you ask ChatGPT about it, and enforcing the shared free usage limits.
• Recipients: Our backend; Groq, which transcribes the audio, with fal.ai and our own speech recognition server as fallbacks; and ChatGPT, which receives the transcript text and its timestamps whenever the transcription result is requested in your conversation, and when you send a message while the transcript is open in the Spokenly workspace.
• Retention: Upload links expire after five minutes. Unlike the built-in cloud models above, our backend temporarily keeps an encoded copy of the audio for playback, the transcript, and the file details. Each expires automatically no later than one hour after it is stored. Shared usage counters, which contain no audio, transcripts, or file details, expire after two days. Standard server logs may record IP addresses for security and troubleshooting.
• Your choices: You decide which recordings to send, and you can disconnect the plugin in ChatGPT at any time. The plugin keeps no account or history to manage. Copies in your ChatGPT conversation are governed by OpenAI's policies and your ChatGPT data controls. For privacy questions, contact admin@spokenly.app.

When using the Spokenly plugin for Claude: The same data, purposes, retention and choices apply, with Claude in place of ChatGPT and without ChatGPT file IDs. Groq transcribes the audio, with fal.ai and our own speech recognition server as fallbacks. The plugin processes recordings you upload in the plugin and public HTTPS links you provide, which our backend uses only to download the audio. Claude receives the transcript text and its timestamps whenever the transcription result is requested in your conversation, and when you send a message while the transcript is open in the Spokenly view. So the Spokenly view can reopen your recording after Claude reloads it, a temporary transcription ID and its expiry time are kept in this browser tab's session storage. Restoring works for up to one hour, expired IDs are discarded when the view reopens, and the browser clears the tab's session storage when you close the tab. Audio and transcript content are not stored there. You can disconnect the plugin in Claude at any time. Copies in your Claude conversation are governed by Anthropic's policies and your Claude data controls.

We make no permanent copy of your recordings on our servers. Apart from the temporary ChatGPT and Claude plugin playback copies described above, any temporary copies in transit are used solely to facilitate transcription. Nonetheless, you should be aware that your audio data is transmitted over the internet and handled by our third-party providers in accordance with their own privacy practices.

3. Chrome Extension

The Spokenly Chrome extension records audio only while you dictate. With built-in cloud models, the audio is handled as described above. If you choose Chrome Speech Recognition, your audio is processed by Chrome's built-in speech recognition, which sends it to Google. It does not pass through Spokenly's servers, and Google's Privacy Policy applies.

The extension's use of data complies with the Chrome Web Store User Data Policy, including the Limited Use requirements.

4. Third-Party Services

We use third-party providers to support transcription and AI text processing features, with the specific provider chosen based on your selected model, feature, and fallback path. These providers may include OpenAI, Deepgram, Soniox, Cartesia, ElevenLabs, Groq, Mistral AI, xAI, Microsoft Azure, Google Cloud, AssemblyAI, OpenRouter, fal.ai, and Cerebras. OpenRouter may route requests to underlying model providers depending on the selected model.

For Spokenly-managed cloud features, we select providers and model configurations with privacy in mind and prefer options that do not use customer data for model training where available. Separately, if you configure your own API key, the app sends audio or text directly from your device to the provider you selected, and it never passes through Spokenly's servers; that provider's own terms and settings apply. While we do not keep your audio data beyond the temporary ChatGPT and Claude plugin playback copies described in section 2, these third-party providers may process and/or store certain data for their own purposes. We recommend reviewing their respective Privacy Policies if you have any concerns about how they handle user data.

The assistant chat on our website sends your messages, the address of the page you are on and your browser language to our server, which forwards them through OpenRouter to an AI model provider listed above to generate an answer. We may read chat questions and answers to improve the assistant and our documentation. Please do not share personal or sensitive information in the chat. Your browser keeps the chat in its local storage until you start a new chat or clear your browser data. If you choose Contact support in the chat, we store your email address, your message and the chat as a support ticket so our team can reply to you.

5. Children's Privacy

Spokenly is not directed toward children, and we do not knowingly collect any information from minors. We encourage parents and guardians to guide minors regarding safe online behaviors.

6. Payment Processing

We do not receive or store your payment card details. Depending on where you purchase Spokenly, payments are handled by one of the following:

• Apple App Store: Apple manages billing and payment processing according to their privacy practices. We use RevenueCat to verify App Store subscription status and sync entitlements.
• Google Play: Google Play handles payments for Android subscriptions. We receive purchase records and subscription status to verify purchases, manage access and prevent fraud. We do not receive or store your payment card details.
• Polar (polar.sh): For website purchases processed by Polar, Polar acts as the Merchant of Record and handles billing, invoicing, and applicable sales taxes. Polar processes your email address and payment details in accordance with their Privacy Policy.
• Paddle (paddle.com): For website purchases processed by Paddle, Paddle acts as the Merchant of Record and handles billing, invoicing, and applicable sales taxes. Paddle processes your email address and payment details in accordance with their Privacy Policy.

7. Data Security

We take industry-standard security measures to protect your data, including the transfer of audio recordings to transcription services and the storage of account information. We strive to keep our service secure, but we cannot guarantee absolute protection of data transmitted across the internet.

8. Changes to This Policy

We may update this policy over time as Spokenly grows and changes. Any revisions will be posted here with an updated "Last Updated" date. We encourage you to check back periodically to stay informed about how we are protecting your data.

9. Contact Us

If you have any questions, concerns, or feedback about this Privacy Policy or Spokenly, please reach out to us at: admin@spokenly.app.