Disclosure: this post links only to my own site, infinityme.dev, and some pages there contain affiliate links. If you sign up through them I may earn a commission at no extra cost to you. Descript doesn't sponsor this post.
Descript's pitch is "edit video like a document": delete a sentence in the transcript and the cut is done. For talking-head videos, tutorials and podcasts it really is faster than scrubbing a timeline. The question I had before paying was more boring: which plan actually covers how much I record?
Descript meters usage in media hours, and the number that matters is not the length of your finished video. So I wrote a ~60-line stdlib Python script that answers it from either a rough estimate or the actual WAV files on your disk.
All plan numbers below come from Descript's official pricing page as summarized in my Descript review (checked October 21, 2026). Descript changes plans fairly often, so confirm on the live page. The script and its example numbers are mine.
The plans, as of October 2026
| Plan | Billed annually | Billed monthly | Media hours / month | AI credits | Max export |
|---|---|---|---|---|---|
| Free | $0 | $0 | 1 hour | 100 (one-time) | 720p |
| Hobbyist | $16/mo | $24/mo | 10 hours | 400 / month | 1080p, no watermark |
| Creator | $24/mo | $35/mo | 30 hours | 800 / month | 4K, no watermark |
| Business | $50/mo | $65/mo | 40 hours | 1,500 / month | 4K, no watermark |
| Enterprise | Custom | Custom | Custom | Custom | 4K |
Prices are per person (per editor) in USD. Three details from the pricing page drive the math:
- Media minutes are minutes of imported or recorded media processed in Descript. That means your raw footage, retakes included, not your final runtime. A 20-minute video cut from 60 minutes of recording costs you 60 minutes.
- There are two limits besides price: media hours and export resolution. Free tops out at 720p and Hobbyist at 1080p, so needing 4K skips you straight to Creator.
- You can buy top-ups for both media minutes and AI credits, so going slightly over isn't fatal, just extra cost.
The script
"""Which Descript plan covers your monthly raw footage? (stdlib only)"""
import argparse, pathlib, wave
# name: (annual $/mo, monthly $/mo, media hours/month, max export height)
PLANS = {
"Free": (0, 0, 1, 720),
"Hobbyist": (16, 24, 10, 1080),
"Creator": (24, 35, 30, 2160),
"Business": (50, 65, 40, 2160),
}
def wav_minutes(folder):
"""Total duration in minutes of every .wav file under folder."""
total = 0.0
for p in sorted(pathlib.Path(folder).rglob("*.wav")):
with wave.open(str(p), "rb") as w:
total += w.getnframes() / w.getframerate()
return total / 60
def check(media_min, need_height=1080, billing="annual"):
rows = []
for name, (annual, monthly, hours, height) in PLANS.items():
price = annual if billing == "annual" else monthly
left = hours * 60 - media_min
ok = left >= 0 and height >= need_height
rows.append((name, price, hours, height, left, ok))
pick = next((r for r in rows if r[5]), None)
return rows, pick
def main(argv=None):
ap = argparse.ArgumentParser(description=__doc__.splitlines()[0])
src = ap.add_mutually_exclusive_group(required=True)
src.add_argument("--raw", type=float, help="raw minutes recorded/imported per video")
src.add_argument("--wav-dir", help="sum .wav durations in this folder instead")
ap.add_argument("--videos", type=int, default=1, help="videos per month (with --raw)")
ap.add_argument("--height", type=int, default=1080, help="export height you need: 720/1080/2160")
ap.add_argument("--billing", choices=["annual", "monthly"], default="annual")
a = ap.parse_args(argv)
media = wav_minutes(a.wav_dir) if a.wav_dir else a.raw * a.videos
rows, pick = check(media, a.height, a.billing)
print(f"Media per month: {media:.0f} min ({media / 60:.1f} h), need {a.height}p, {a.billing} billing")
for name, price, hours, height, left, ok in rows:
why = "fits" if ok else ("export too small" if height < a.height else "out of media hours")
print(f" {name:<9} ${price:>3}/mo {hours:>2} h {height:>4}p {left:+7.0f} min {why}")
if pick:
print(f"Cheapest plan that fits: {pick[0]} (${pick[1]}/mo, {pick[4]:.0f} min spare)")
else:
print("Nothing fits: buy media-minute top-ups or ask about Enterprise.")
return pick[0] if pick else None
if __name__ == "__main__":
main()
Plans are checked cheapest first, so the first one that fits both limits is the answer. --wav-dir uses the stdlib wave module to total the real length of last month's recordings, which beats guessing.
Three channels, three answers
A weekly podcast: 90 minutes recorded per episode, 4 episodes a month, 1080p is fine.
$ python descript_fit.py --raw 90 --videos 4
Media per month: 360 min (6.0 h), need 1080p, annual billing
Free $ 0/mo 1 h 720p -300 min export too small
Hobbyist $ 16/mo 10 h 1080p +240 min fits
Creator $ 24/mo 30 h 2160p +1440 min fits
Business $ 50/mo 40 h 2160p +2040 min fits
Cheapest plan that fits: Hobbyist ($16/mo, 240 min spare)
Twice-a-week YouTube in 4K: 75 raw minutes per video, 8 videos a month. Hobbyist has exactly enough media hours, but it can't export 4K.
$ python descript_fit.py --raw 75 --videos 8 --height 2160
Media per month: 600 min (10.0 h), need 2160p, annual billing
Free $ 0/mo 1 h 720p -540 min export too small
Hobbyist $ 16/mo 10 h 1080p +0 min export too small
Creator $ 24/mo 30 h 2160p +1200 min fits
Business $ 50/mo 40 h 2160p +1800 min fits
Cheapest plan that fits: Creator ($24/mo, 1200 min spare)
A long-form interview show on monthly billing: 3 hours recorded per episode, 4 episodes. The finished episodes might be 4 hours total, but the raw footage is 12, which pushes past Hobbyist.
$ python descript_fit.py --raw 180 --videos 4 --billing monthly
Media per month: 720 min (12.0 h), need 1080p, monthly billing
Free $ 0/mo 1 h 720p -660 min export too small
Hobbyist $ 24/mo 10 h 1080p -120 min out of media hours
Creator $ 35/mo 30 h 2160p +1080 min fits
Business $ 65/mo 40 h 2160p +1680 min fits
Cheapest plan that fits: Creator ($35/mo, 1080 min spare)
That last case is the one people miss: they budget by final runtime and run out mid-month. Being 2 hours over Hobbyist might be cheaper to cover with a top-up than to upgrade; check the current top-up price before deciding.
I keep descript_fit.py next to a 7-case unittest file covering the exact-10-hour boundary, the 4K rule, monthly prices and the WAV-folder mode (it writes three 2-minute silent WAVs into a temp dir and checks they total 6 minutes).
What the script doesn't measure
- AI credits. Green Screen, Eye Contact, automatic clips and Underlord edits all consume credits (400, 800 and 1,500 a month on the paid tiers). If you lean on those, credits can run out before media hours do.
- Seats. Prices are per editor. Creator can scale to a team of 3 and Business to 5, billed separately.
-
Annual vs monthly. Descript says annual billing saves up to 35%. Starting monthly for a month, then running the script on your real
--wav-dir, is a cheap way to avoid committing to the wrong tier. - Free is a test drive. 60 minutes a month and 720p export are enough to see whether transcript editing clicks for you, not to run a channel.
Takeaway
For any minute-metered creator tool, find out what the meter actually counts. For Descript it's raw media in, not polished video out, so measure your recordings, not your uploads.
Features, pros and cons and who Descript is (and isn't) for are in the full Descript review on infinityme.dev.
Top comments (0)