A meditation's duration comes from its speech, its pauses and the automatic gaps between them. Open Meadow estimates it from the script before any audio exists, and can adjust pauses to reach a target duration.
How Open Meadow estimates duration
The estimate adds up three parts:
- Speech: Open Meadow estimates how long each passage takes to read aloud, using a model trained on real voice audio. Non-Latin scripts, such as Cyrillic, Persian or Chinese, have their own calibration.
- Pauses: each pause counts at its written length.
- Automatic gaps: about 2 seconds between paragraphs and 3 seconds before each section. See Pauses.
Titles and section names are not spoken, so they do not count.
The script editor shows the estimated duration of the script as written. Meditation cards in My Library show the length after the target duration and spaciousness are applied, so the two can differ. The real audio can differ slightly, because voices speak at different speeds.
Target durations
A meditation can have a target duration, for example when you ask Omi for "a 10-minute meditation" or choose a duration in the Instant Player.
- Omi plans the speech and pauses together to fit the target.
- When the script is complete, Open Meadow scales the standard and open pauses so the duration matches the target.
Once speech audio exists, the scaling uses the real length of that audio, so the pauses also make up for a voice that speaks faster or slower than estimated. Speech and exact pauses do not change. A script with only exact pauses, or with very little silence, cannot always reach its target. When Omi preserves a script as written, the target is not guaranteed.
The spaciousness slider adjusts pauses further while you listen, so the playing length can differ from the target.
Writing to a duration in another AI tool
Another chatbot can write a script of a given length with this rule of thumb:
Duration in seconds ≈ spoken words ÷ 1.5 + total pause seconds
For example, a script with 300 spoken words and 400 seconds of pauses lasts about 300 ÷ 1.5 + 400 = 600 seconds, or 10 minutes.
Without a check, scripts from general-purpose chatbots often run 1.5 to 2 times the requested length, because the model estimates speech loosely and overlooks the pauses. These instructions help:
- Aim for the requested duration by planning the spoken guidance and pauses together.
- Count only spoken text and pauses. Titles and section headings do not count.
- Allow about 1 second for every 1.5 spoken words, then add every pause as written.
The Meditation script format guide includes these instructions, so a prompt like this works:
Follow the Open Meadow script format and its duration-checking
instructions. Check your calculation before giving me the final script.
Aim for a total duration of 10 minutes.
Chatbots follow the calculation reliably only with reasoning or "thinking" turned on. Fast settings, such as ChatGPT's Instant mode, may skip it. When the script reaches Open Meadow, an adapted script can be refined further toward the target.