Meeting Notes Summarizer

Q: Will this tool invent an action-item owner or due date?

It is explicitly instructed not to. Missing owners and dates are stored as unknown, and every action item must cite a transcript segment for review.

Q: Can I use an audio recording instead of a transcript?

Yes. The P1 input enhancement adds audio upload and microphone recording, then uses the single shared local Whisper runtime before creating notes on this page.

Q: Is meeting content uploaded?

No. Transcript processing, model inference, source references, session storage, and exports stay in your browser.

Q: Does it identify different speakers automatically?

No speaker diarization is promised in this MVP. Existing speaker labels in the transcript are preserved as evidence, but unlabeled audio may produce an unlabeled transcript.

Turn a pasted or imported meeting transcript into source-linked minutes, decisions, risks, unresolved questions, and action items. Unknown owners and dates stay explicitly unknown, and all processing remains in your browser.

Saved sessions

Transcribe a recording

451 / 1000000

1 transcript segments

Owners and due dates are kept as unknown unless the transcript states them explicitly. Audio upload and microphone transcription are added here through the shared Whisper runtime.

Local AI runtime

Checking this device

BackendWASM

Device profile—

Adaptive limits—

Writing model

Qwen 0.5B · standard

Transcription model

Whisper Tiny · fast

Files, prompts, and results stay in this browser. Browser storage can still be cleared by you, private mode, or browser cleanup.

Private local meeting processing

The transcript, prompts, notes, action items, and source references remain in your browser. Later audio transcription uses the same global local Whisper runtime and does not upload recordings.

Current limitations

The model can miss or misclassify ambiguous discussion, so decisions and action items should be checked against their cited transcript segments.
Unknown owners and due dates remain unknown; the tool will not infer responsibility from role or speaking order.
This MVP does not promise speaker diarization or perfectly real-time audio transcription.
Model weights are cached by version when browser storage allows it, but site-data clearing or eviction can require another download.

How the meeting notes workflow works

Provide a transcript

Paste or import text now; the same page also accepts locally transcribed audio through the shared Whisper enhancement.

Extract structured evidence

The global local model extracts minutes, decisions, risks, questions, and actions from bounded transcript segments.

Validate, review, and export

Every item must cite a real source segment. Review unknown fields and export notes as Markdown or TXT and actions as CSV.