SimplyArticulate

Privacy and recordings

Updated October 31, 2026. Questions and privacy requests: simplyarticulate11@gmail.com.

Recording on your device

Microphone permission is requested when you check the microphone or start recording. Recording and replay do not by themselves request AI analysis. The browser can keep local audio recovery copies, settings, introduction preferences and account-specific completion markers. Closing a tab does not necessarily remove browser storage. Download important audio before clearing site data.

When you request AI feedback

Your recording is uploaded for processing. AssemblyAI transcribes audio. Google Gemini receives transcript and practice context for coaching; factual claims may be checked with Google Search. The hosted backend uses Supabase for authentication, account data, private storage and analysis jobs. These providers and hosting services process the requests needed to operate the service, which can include technical connection and error information. Avoid recording confidential information or other people without permission.

Your coaching plan, grades and live hints

When you ask My plan to build your next practice session, a few short quotes from your recent takes, with their scores and speaking measurements, are sent to Google Gemini so it can choose drills and goals. Whole transcripts are not sent for this. Grades and feedback are written by AI from the transcript, so they can be wrong. Very short or hard-to-transcribe takes are not graded. The live pace hint during recording is worked out on your device from loudness alone; nothing is transcribed or sent while you speak.

Saving and sharing

Scores and the next practice focus are kept with progress. Practice checkmarks and notes you choose to save are kept privately with your account progress. Full transcripts and coaching from practice and coaching takes are kept privately with your account by default, so your coaching plan can use them. You can turn this off for any take, or remove saved feedback later. For daily challenges, keeping them is a separate choice. Saving audio to your account is another separate action. A public daily-challenge entry can expose your public username and score; publishing your transcript requires its own choice. Other visitors can read material you choose to publish.

Retention and deletion

You can delete your account and its saved data yourself in Settings, or ask us to do it. Temporary analysis audio and provider transcripts are scheduled for cleanup after processing. Background failures can delay cleanup while it retries. Saved account content remains available until you remove it or request account deletion. Removing a saved feedback report also removes its practice checkmarks and note; account deletion removes these annotations with the rest of your progress. Account deletion runs in stages and may need to be resumed if a stage fails. Local recovery data on other devices is not remotely erased. Operational cleanup records and service-provider logs can outlast the visible content; contact us for questions about a specific request.

Your choices

You can use recording and replay without signing in, decline AI analysis, download audio, remove saved feedback or recordings, and request account deletion in Settings. For access, correction, export or deletion questions, email simplyarticulate11@gmail.com. We may need to verify account ownership before disclosing or removing account data.

Advertising

The current application does not load the AdSense delivery script or show live ads. A Google AdSense account meta tag is used for site verification. If Google advertising is activated, this notice and any applicable consent controls must be updated before ad requests begin. Google explains advertising-cookie controls in its advertising policy.

Earlier debate rounds

Opponents in earlier debate rounds were labeled AI characters or consenting past performances. Other players cannot access your audio or transcript. To prevent repeated prompts, your account retains the identifiers and first-seen times of topics you have faced until account deletion, even after temporary speech records expire. Bounded, private round-stage events help us measure reliability and repeat use; they contain no audio, transcripts, typed questions, email or IP addresses and are removed after 30 days. You can withdraw legacy argument-reuse consent in Settings. New argument-reuse opt-ins are disabled.

Billing

Paddle acts as the Merchant of Record for Pro subscriptions and processes payment and billing details. We use Paddle customer, subscription and transaction identifiers, billing status and relevant purchase information to manage account access and billing support. We do not store full payment-card numbers. Your purchase email and any billing information you enter at checkout are handled by Paddle under its Privacy Policy.

Spoken AI responses

When Azure voice playback is enabled, Microsoft Azure Speech receives the generated AI response text to create audio. That response may quote information you supplied; avoid including sensitive details in practice. Your microphone recording is not sent to Azure for this feature. Generated playback audio stays in page memory and is released when the response is closed. Private usage counters and request identifiers prevent duplicate generation and enforce voice limits; account-linked counters are deleted with your account and aged out after 100 days during subsequent voice use. Aggregate monthly usage totals do not identify users. If Azure is unavailable or its app allowance is reached, playback falls back to an available device voice.

Service providers

Provider processing also depends on the service configuration and agreements in use. See Microsoft, AssemblyAI, Google, Supabase and Netlify for their published privacy information.