SpeakSignal Offline

Practice American English pronunciation one sentence at a time, on your phone and offline. Pick one of 20 sound groups from the “Sound to practice” dropdown, then choose your topic and Easy, Medium or Hard level. Topics include Everyday life, Work & meetings, Interviews, Small talk, Travel, Technology and Presenting.

Tap Listen to hear the sentence using an installed offline voice. Record a take of up to 120 seconds, then compare the expected and heard sounds in the report. The on-device sound model gives a verdict for your chosen group and lists every other difference it heard under “Also heard.” These are model estimates: the check can miss a difference or mishear a sound, and some groups are only partly checked.

The “Sounds heard” line shows the estimated sounds in phonetic symbols (IPA), with differences marked in red and underlined. “Full report: all sounds and mismatches” opens automatically and groups the details by word. It shows the expected and heard sounds, marks uncertainty with a question mark, and identifies words that were not heard or could not be checked.

Choose from groups including TH, V / F, American R, SH / CH, S / Z, m / n / ng, food / foot, and Pete / pit / pet / Pat / pot. Each group tells you whether it is checked by the model or partly checked, with a note explaining its limits where needed.

Practice sentences are included on the phone. You can also type or paste your own English passage, up to 80 words and 800 characters. On compatible devices, optional on-device AI can write new sentences for your chosen topic and sound, including a custom topic. An initial AI model or voice download may need internet; practice with the installed sound model, library and offline voice works without a connection.

Recordings and sound analysis stay on your phone. Recent practice summaries are stored locally and can be deleted in Settings. No account or subscription is required. This edition is intended for adults.