Questions about speaking practice, privacy and how nthoughts works
Answers to the most common questions from learners. More detail on how we handle voice data.
Frequently Asked Questions
Common Questions
Clear answers on our speaking drills, local data sovereignty, supported accents, and design philosophy.
You upload any standard PowerPoint (.pptx) deck. nthoughts extracts the slide structure, metrics, tables, and speaker notes. You rehearse slide-by-slide with real-time speech transcription and a live pace gauge. Afterward, you receive slide-aware AI feedback showing which key metrics you covered, continuity transition bridges between slides, 4-way adaptive sentence rewrites, and simulated stakeholder Q&A questions.
Yes. Uploaded slide decks are parsed securely in memory for your session only. Raw audio is never permanently stored, and your slides and transcripts are private to your user account and never shared with third parties or used for public AI model training.
Yes. nthoughts is free during launch — no credit card required. There is a daily fair-use limit on AI-backed requests. Paid plans are not available yet.
We are planning a Pro plan with higher limits, but nothing is on sale yet and nothing changes for you today. If you are interested in team use, contact us and tell us what your team needs.
Maya uses high-speed streaming speech-to-text to transcribe your spoken voice in real time. The nthoughts linguistic engine analyses your syntax, preposition usage, and vocabulary while you speak, isolates mistakes without interrupting your conversational flow, and replies using natural ElevenLabs synthetic speech.
Yes. In the session controls, you can adjust Maya's speaking rate from 0.8x (for clearer comprehension) up to 1.2x (for native-paced conversational challenge). You can also switch between different voice models and accent styles.
nthoughts is built specifically for adult professionals who have intermediate English comprehension (CEFR A2 to C1) but freeze or stumble when speaking under pressure. While beginner-friendly prompts exist, the workspace delivers the highest value to those who already read or write English but lack oral fluency.
Yes. nthoughts includes dedicated practice modules for behavioral interview questions, technical architectural explanations, and spontaneous responses required in IELTS and TOEFL speaking tests.
The Translation Speaking Game currently supports Tamil, Telugu, Hindi, and Malayalam. You see a sentence prompt in your native script and must speak the accurate, idiomatic English translation out loud before the timer expires.
Traditional apps rely on multiple-choice taps, spelling cards, and artificial gamification like XP and streaks. nthoughts is 100% focused on active oral production: you speak continuously into your microphone, receive instant grammar isolation, and reinforce only your real spoken mistakes.
ELSA Speak specialises in phoneme-level pronunciation drills (reading scripted sentences aloud). nthoughts emphasises spontaneous conversational synthesis with Maya, instant grammatical repair, and translation games that break your habit of translating word-by-word in your head.
Yes. You can cancel your subscription at any time with a single click inside Settings. You will retain Pro access until the end of your current billing period, with zero hidden cancellation penalties.
Research in cognitive psychology demonstrates that streak mechanics create artificial performance anxiety and encourage vanity usage (rushing through 2-minute superficial exercises just to keep a counter alive). nthoughts is a calm, streak-free workspace designed for deep, focused speaking practice when you are genuinely alert.
Standard spaced repetition apps feed you pre-made vocabulary flashcards. nthoughts creates personal Leitner memory cards exclusively from errors you made in spoken conversations with Maya. If you miss a preposition like 'discuss about', it reappears in your review queue at scientifically calculated intervals until mastered.
Reading and listening rely on passive recognition, which uses low cognitive effort. Speaking requires active retrieval, grammatical assembly, and physical motor coordination under real-time social pressure. nthoughts trains oral automaticity so your brain retrieves sentence structures without conscious hesitation.
Fossilisation occurs when non-standard grammar habits (such as omitting articles or using present continuous for habitual action) become permanent through repetitive uncorrected practice. Maya highlights the exact phrase within 2 seconds of speaking, enabling immediate neural self-correction.
Adult bilinguals often think in their primary language and translate word-by-word, leading to awkward phrasing. By presenting rapid native prompts in Tamil, Telugu, Hindi, or Malayalam, the game builds cognitive bridges directly between concept and spoken English phrasing, bypassing literal translation.
No. We do not store raw audio files on our servers. Your voice is transcribed in real-time browser memory. Short audio snippets sent for pronunciation scoring are processed via ephemeral TLS connections and deleted immediately after scoring.
No. We have strict commercial zero-data-retention agreements with our infrastructure and AI model providers. Your practice speech and transcripts are never used to train public or foundation models.
No. If your company provides your nthoughts seat via a remote team licence, administrators only see aggregated usage metrics (such as active seats and total practice minutes). Your conversation logs, error history, and transcripts remain strictly confidential to your account.
You can permanently purge your account at any time directly from the Settings page. This deletes your user profile, saved mistake history, and session analytics from our databases immediately and irreversibly.
All account metadata and encrypted mistake logs are stored in SOC 2 Type II compliant cloud facilities. All data in transit is encrypted using TLS 1.3, and all database volumes are encrypted at rest with AES-256.
Yes. nthoughts is fully responsive and optimised for mobile web browsers, including Safari on iOS and Chrome on Android. Full microphone and audio synthesis APIs run natively without requiring an App Store download.
No. nthoughts is a progressive web application. You can access your full speaking studio from any modern desktop, tablet, or phone browser without downloading or installing third-party binaries.
No. The built-in microphone on modern laptops and smartphones is completely sufficient. However, if you are practising in an open office or noisy coffee shop, a standard wired or Bluetooth headset with background noise suppression is recommended for optimal speech recognition.
Maya's acoustic models are trained on diverse international accents, including South Asian, East Asian, Latin American, European, British, American, and Australian English. You do not need to alter your natural accent to be accurately understood.
Yes. Our transcription engine is specifically calibrated for Indian English phonology, accommodating distinct syllable-timed rhythms, retroflex consonants, and cadence patterns common among Tamil, Telugu, Hindi, and Malayalam speakers.
Cached Leitner vocabulary cards can be reviewed in your browser without an active connection. However, real-time voice conversations with Maya, live pronunciation analysis, and Translation Speaking rounds require an active internet connection to communicate with our AI engine.