Preferences
Usage
See your dictation activity, compare model processing times, and understand the profiles you use most.
In the appSettings → Usage
On this page
Open your usage dashboard
Open the menu-bar panel, choose Settings → Usage, and see how much work your speech models and rewrite profiles have handled. The dashboard groups activity by model and profile, with word counts, token counts, and processing times.
Models appear after they have run. Usage is collected on this Mac and kept between launches, independently of transcript history.
Choose a period
Use the switch at the top to choose Last 7 days, This month, or All time. This month is selected when you open the pane.
- Last 7 days: today and the previous six calendar days.
- This month: activity since the start of the current calendar month.
- All time: all retained usage since tracking began or the counters were reset.
The selected period applies to the summary, model breakdowns, and profile breakdown. A model with no activity in that period is omitted.
Read the summary
The headline figures are counted in words.
| Figure | What it counts |
|---|---|
| Consumed | Transcript words read by rewrite models. Transcription takes audio as input, so it adds no input words. |
| Produced | Words generated across transcription and rewriting. |
| Dictations | Dictations with words inserted into the destination app, counted once per dictation. |
| Words typed | Words inserted into the destination app. Clipboard-only results are included in model activity, but do not add to Words typed or Dictations. |
Compare activity by model
Turning speech into text shows the transcription models you used. Rewriting before it is typed shows the models that prepared your transcripts. Each row identifies the model and its location or provider.
| Figure | How to read it |
|---|---|
| Runs | The number of recorded runs for that model or processing step. |
| Listened to | The duration of audio processed by transcription models. |
| Spent | The total processing time for the selected period. |
| Words in / out | Transcript words read and text words produced. Speech models show audio as their input. |
| Tokens in / out | Model token counts, where available. See the explanation below. |
| Average | The average processing time per run. |
| × real time | Audio duration divided by transcription time. Processing 60 seconds of audio in 15 seconds gives 4× real time. |
Transcription bars compare the amount of audio each model handled. Rewrite bars compare total token activity, with separate colors for input and output. The busiest model by that measure sets the full bar length.
Use the timing figures to compare models with your usual dictations. Local model timings exclude initial model loading; provider timings include the network request.
Understand words and tokens
Words describe the text you read. Tokens are the smaller units a model processes, so word and token counts differ. Rewrite input tokens can also include your profile instructions, while Words in counts the transcript itself.
MicMyDay uses token counts reported by supported models. A value marked ≈ is estimated from the text, such as when using Apple built-in rewriting. Some engines do not supply token counts; a zero in that column can mean no token measurement was provided.
Use word counts to understand the amount of text, and token counts to understand model activity.
See which profiles you use
Which profile asked for it groups rewrite activity by profile. Each row shows runs, words and tokens in and out, and average processing time.
This is another view of the same rewrites shown under models. Use it to see which profiles handle most of your writing, such as email preparation, notes, or agent prompts. The section appears once a profile has recorded rewrite activity in the selected period.
Keep or reset your counters
Usage stores numeric totals and model and profile information locally. It does not store transcript text or recordings. Counters continue accumulating when transcript history is off.
- In Settings → Usage, find These counters.
- Choose Reset counters.
- Confirm with Reset them, really, or choose Keep them to cancel.
Resetting clears usage for all periods and shows when counting restarted. It leaves your transcripts, models, profiles, and settings intact. Clearing transcript history does not reset usage counters.
For the other data controls, see Privacy & your data.