Weekly 1-hour show: yearly budget for transcript and clips

A weekly one-hour show comes to $114.40 a year at Sume's public per-job and per-minute rates for a transcript plus five captioned vertical clips each week ($2.20 a week), before the inspect compute line. With ten clips a week it is $197.60 a year. Caption jobs are the largest line, not the transcript.

The week is priced like this: a 60-minute transcript at $0.01 per audio minute ($0.60), then for every clip a $0.02 trim, a $0.20 caption job and a one-minute Timeline render at $0.10. Rates are Sume billable amounts read 2026-10-03.

The yearly table

Clip counts are our planning scenarios. Each clip is assumed to be 60 seconds or shorter, so it is one caption job and one Timeline minute. The transcript line is the same in every row because the show length does not change.

What the numbers say

The transcript is $0.60 of a $2.20 week at five clips, which is 27 percent. Captions are $1.00, or 45 percent. So if you want to reduce the yearly bill, reduce the number of captioned clips or burn the captions in one step on fewer, longer pieces.

Doubling the show length from 60 to 120 minutes only adds $0.60 a week; adding five more clips adds $1.60. Clip count, not show length, is the lever.

Per clip and per hour of show

Spread over five clips, a week is $2.20 divided by 5, or $0.44 per finished clip, transcript share included. Over a year that is 260 clips for $114.40. The transcripts alone for 52 hours are $31.20.

If you only caption the three best clips, the week drops to $1.56 and the year to $81.12, as the first row shows. Cutting captions to one in two clips saves more than any change to the transcript, because the caption job is twenty times the per-minute transcript rate.

Where the 1,800 second cap bites

Video trim and Video inspect both refuse a source longer than 1,800 seconds. A one-hour show therefore has to arrive as two files of up to 30 minutes each. Cut your clips from those files. That is a workflow detail rather than a cost, since the rate does not change.

Inspect also reserves its Modal compute ceiling and settles at container seconds, which is a separate line from the transcript rate. The weekly figures here are the public per-job and per-minute rates; compute for inspect shows on the receipt.

What Sume does and does not do

Sume returns the transcript text with word times and cuts the clips you name; it does not choose the best moments for you. Choosing the moments is a human or agent step before you call trim. Sume also will not publish the clips. Confirm rates in GET /v1/catalog.

• Transcript: $0.01 per audio minute.
• Trim: $0.02 per job.
• Captions: $0.20 per job up to 60 seconds.
• Timeline: $0.10 per ceil(output minute).

Rerun the arithmetic

The same line items as a short script, so you can change a quantity and see the total. It is plain arithmetic on the rates in this post; it does not call the Sume API.

# Sume billable USD rates, read 2026-10-03. Edit the quantities.
lines = [
    ('Transcript minutes per week x 52', 3120, 0.01),
    ('Trim jobs, 5 a week x 52', 260, 0.02),
    ('Caption jobs', 260, 0.20),
    ('Timeline minutes', 260, 0.10),
]
total = 0.0
for label, qty, unit in lines:
    cost = qty * unit
    total += cost
    print(f"{label:<46}{qty:>8g} x {unit:<10g} = ${cost:,.2f}")
print(f"{'Total':<46}{'':>8}   {'':<10}   ${total:,.2f}")