Gling silence removal at $20 for 10 hours vs a Sume cut list

Gling Plus is cheaper per talk if you use its whole allowance, and Sume is cheaper if you edit only a few talks a month. Gling's pricing page (read 2026-10-03) lists Plus at $20 a month for 10 hours of AI-edited media, which is $2.00 per hour of raw recording billed monthly. A Sume cut list for a 10-minute talk edited to a 3-minute Reel runs about $0.41, and 15 minutes of raw talk edited down to 6 minutes runs $0.76 in the calculator below.

The comparison matters this week because Instagram now lets Reels run to 3 minutes and its iPhone First Draft feature trims pauses out of multiple clips into a rough cut, as Metricool's Instagram news page reports (read 2026-10-03). Everyone who edits talking-head video is looking at pause removal again, and the tools split into two camps: an app that decides the cuts for you, and an API where you supply them.

What does Gling charge?

All numbers below are from the pricing page as fetched on 2026-10-03. Gling lists a Free plan with 1 hour of AI-edited media a month, then Plus, Pro and Elite. Paid plans include bad-take removal, automatic silence removal, filler-word removal, AI captions, auto framing and multicam editing. Paid plans export without a watermark and can export XML to Premiere, Final Cut Pro or Resolve. The page states no API and no maximum video length, so this post does not claim either way.

Dividing the monthly fee by the included hours gives the effective hourly rate. That is a ceiling on what you pay if you use every hour; unused hours do not lower the price.

What does the same edit cost on Sume?

The Sume docs list no silence detector and no filler-word remover, so the first job is a transcript. Video inspect with transcribe: true adds speech-to-text at $0.01 per audio minute on top of the probe's Modal compute and, with segmentation.mode: "sentence", returns gapless sentence segments plus word times. A few lines of your own code turn the gaps between words into a keep list; that is the part Gling does for you and Sume does not.

The keep list then drives two steps. Audio detach pulls the talk's audio into a durable file for $0.01. Timeline 1.0 renders the result at $0.10 per output minute, rounded up. The audio spine can be built from up to 20 audio.parts[] slices, each with a source_in and a duration, and every video[] slot carries the matching source_in, so the cuts happen inside the render and you do not need a separate trim job per keep. Past 20 keeps, merge the audio once with timeline audio for $0.01 per job.

def sume_cost(raw_min, kept_min, slices=1):
    """One talk: transcript, one audio detach, timeline_audio per 20 slices, render."""
    import math
    stt = 0.01 * math.ceil(raw_min)
    detach = 0.01
    split = 0.01 if slices > 20 else 0.0   # merge once past 20 slices
    render = 0.10 * math.ceil(kept_min)
    return round(stt + detach + split + render, 2)

for raw, kept in ((10, 3), (10, 6), (15, 6), (15, 9)):
    print(raw, "min raw ->", kept, "min kept:", sume_cost(raw, kept), "USD")

Worked examples

Take a 10-minute talk you want as a 3-minute Reel. On Sume that is 10 transcript minutes ($0.10), one audio detach ($0.01) and a 3-minute render ($0.30), for about $0.41, before inspect's small compute charge. The same talk through Gling Plus, if you used every included hour, is about $0.33 of the monthly fee. Gling's share is eight cents lower per talk, but you pay the $20 whether you use it or not.

Scale it to a 15-minute talk edited to 6 minutes, 900 seconds being the most one audio detach will output. Sume is 15 transcript minutes ($0.15), one detach ($0.01) and a 6-minute render ($0.60), so $0.76, against $0.50 of a Plus hour. Longer sources need one detach range per 900 seconds, and the output side still bills only the minutes you keep, so the gap narrows as the edit gets shorter.

Against Gling Plus the break-even is about 49 ten-minute talks a month ($20 divided by $0.41); below that Sume costs less in dollars, above it the flat allowance does. Sume's render line is $0.10 per finished minute, so a Reel of 3 minutes or less costs $0.30 or less to assemble no matter how long the source was. Gling's line is per hour of source.

What Sume does and does not do here

• Sume does not decide where the pauses are. You derive cuts from the transcript's word times.
• Sume does not offer a visual editor. The cut list is JSON, so it can be reviewed in a pull request and re-run with an Idempotency-Key when a number changes.
• Sume does render the exact keep list you give it: one MP4 per timeline job, up to 1800 seconds of output, with fades and a music bed if you add them.
• Sume does not export XML to Premiere or Resolve. If your editor wants a project file, Gling is the better fit.

What to do

If one person edits a few long recordings a week and likes reviewing cuts in an editor, pay for Gling and stop reading. If you publish many short Reels from long sources, or need the edit to run unattended after an upload, run the transcript through Sume once, write the keep list in code, and cap each render at 180 seconds so a Reel never costs more than $0.30 to assemble. Price one sample first with the unbilled plan call on Timeline, then run the batch.