MemoPen

Meetings, lectures and interviews tend to deliver their most important lines when no one is writing. MemoPen aims to cover those moments without drawing attention: from the outside it is a regular ballpoint pen, so it can sit on a conference table or lecture desk without looking like a gadget. Inside, the maker says, recordings are passed to ChatGPT-5 to be transcribed, organized and queried, cutting out the need to replay and re-summarize audio.

What the AI layer does

According to Tailbone, MemoPen automatically distinguishes between speakers, tagging each statement so a transcript can be sorted by person. It currently supports up to 5 speakers, with an upgrade to more planned. Users can also ask questions of a recording directly, such as the conclusion of a meeting or the tasks that came out of it, and receive an answer generated from the content rather than scrubbing through hours of audio.

The maker claims high recognition accuracy, including technical jargon, industry terms and proper nouns, and support for more than 100 languages including Japanese, English, Chinese and Korean, with the list set to grow through updates. Recordings can be laid out automatically as mind maps, and 31 dedicated structure templates cover scenarios such as meetings, interviews, lectures, research, brainstorming and project management, so a meeting becomes minutes and an interview becomes a Q&A.

Hardware

• Size: 148×11mm; weight: 28g
• 32GB internal storage
• 260mAh battery, up to 30 hours of continuous recording (per the maker)
• Type-C charging
• Built-in OTG connector for direct connection to OTG-capable Android or iOS devices
• Dedicated earphones included for private playback from the pen
• Status lamp showing recording, standby and charging, designed to be visible mainly to the user

Service plans

The maker says activating an account grants one year of Max membership, unlocking AI transcription, 31 AI summary templates and 118-language support. After that, a free Starter plan offers 400 minutes of AI transcription per month. Transcription time is consumed only once per recording; speaker identification, Q&A, summaries and mind maps run on the existing transcript without using further minutes. Tailbone also says the AI features will continue to receive updates.