Recorder Summarize

Why It Exists

Users need to capture and understand audio content without manual transcription. Privacy concerns make cloud-based transcription undesirable for sensitive conversations.

How It Works

Recorder Summarize is Google's on-device speech-to-text transcription and summarization tool powered by Gemini Nano. It records audio, converts speech to text with speaker identification, and generates concise AI summaries highlighting key points. All processing occurs locally on supported Pixel phones using the Tensor NPU, ensuring privacy. The system supports multiple languages, identifies different speakers, and creates searchable transcripts. It is particularly useful for meetings, lectures, interviews, and personal voice memos.

Everyday Use Cases

• Meeting minutes
• Lecture notes
• Interview transcription
• Personal voice memos
• Podcast creation
• Language learning
• Research recording

User Workflow

• User starts recording in the Recorder app
• Audio is captured with speaker identification
• Gemini Nano transcribes speech to text on-device
• AI generates a summary of key topics and action items
• Transcript and summary are saved and searchable

AI Processing Flow

Audio is captured by the device microphones and processed by on-device Gemini Nano speech recognition models. Speaker diarization separates different voices using the Tensor NPU. The transcribed text is analyzed by a summarization model that identifies key topics, decisions, and action items. All processing is on-device with no data upload. Transcripts are searchable and sync via Google account.

Inputs / Outputs

Inputs:

• Audio recording
• Speaker identification
• Language settings
• Timestamp data
• Device context
• Ambient noise profile

Outputs:

• Full text transcript
• Speaker-labeled sections
• Key topic summary
• Action item list
• Searchable content
• Exportable text file

Known Limitations

• Accuracy varies with audio quality and accents
• Supports 10 languages at launch
• Requires Google account for cross-device sync
• Background noise reduces accuracy
• May struggle with overlapping speakers

Unsupported Scenarios

• Music recording transcription
• Professional broadcast quality
• Real-time translation
• Non-Pixel devices
• Live streaming transcription

Performance Notes

Real-time transcription with under 200ms latency; Summary generation in under 5 seconds after recording stops; NPU power consumption under 5 percent per hour of recording; Continuous recording for up to 8 hours on a single charge

Available On

Shows where this feature is available and how its AI processing works on each platform. Availability may vary by device.

Research Status

Confidence and verification reflect how complete documentation is. Fields may show Not assessed or Not yet verified while research is ongoing - this flags gaps, not product deficiencies.

Research Notes

Available on Pixel 8 and newer; Requires Google Tensor G3 or later; Introduced with Pixel 8 Pro 2023; Uses Gemini Nano 1 speech model; Supports 10+ languages; End-to-end encryption option

Details

Capability Mapping

This feature maps to canonical capability family: Text Summarization.

Canonical capability: Text Summarization.

• Text Summarization - Condenses text, messages, documents, or conversations into a shorter summary.

Grounded in 8 device(s) on Gemini Nano.

Related Features

• AI Recorder - Cross-platform equivalent