Meeting Maid documentation

Keep the meeting. Lose the busywork.

Set up two-sided capture, choose where transcription and note generation run, connect calendar context, and publish durable Markdown without giving up control of the source record.

Set up Meeting Maid Solve a problem

Start with the job in front of you

01

Install and record

Grant only the permissions you need and complete a first two-sided recording.

Start setup → 02

Connect your calendar

See today's meetings, join safely, and attach invitations to recordings.

Add context → 03

Choose transcription

Compare WhisperKit, Apple Speech, AWS streaming, and AWS S3 batch processing.

Compare engines → 04

Generate useful notes

Select model roles, tune prompts, name speakers, and preserve every version.

Configure notes → 05

Publish your way

Share individual artifacts or file notes into a plain folder or Obsidian vault.

Set up export → 06

Understand data flow

Know what stays on the Mac and exactly what each optional provider receives.

Review privacy →

A record you can correct

Meeting Maid keeps source audio, timestamped transcript segments, speaker decisions, calendar context, editable title and classification metadata, and versioned Markdown notes connected. You can revise each layer without overwriting the original detection or an earlier note version.

Capture

Microphone + call audio

Record both sides, monitor each level, mute safely, or import existing audio and video.

Review

Transcript + speakers

Edit individual lines, reassign voices, merge labels, undo merges, and reset detection.

Create

Prompts + note versions

Regenerate with another prompt or model while keeping prior results available.

Own

Markdown + portable files

Export audio, transcript, current notes, a complete ZIP, or automatically published copies.

Local and cloud are separate choices

Transcription, speaker analysis, search, and note generation do not share one global privacy switch. Choose the processing path for each capability.

Capability Local option Configured service option
Transcription WhisperKit or Apple Speech AWS Transcribe
Speaker turns FluidAudio AWS native speaker labels
Meeting search Apple English sentence embedding + lexical search None
Notes Local Ollama server Amazon Bedrock or remote Ollama
Start with the privacy guide

The privacy and data reference maps content, storage, provider boundaries, retention controls, and credential handling before you choose a setup.