Audit 100 competitor videos on Sume: $30 analysis plus $20 captions
Video analysis is $0.30 per accepted video on Sume, standalone captions $0.20. 100 videos cost $30, $20, or $50 for both; a trending search is $0.10 a call.

Auditing 100 competitor videos costs $30.00 for analyses at $0.30 each, $20.00 for standalone captions at $0.20 each, or $50.00 for both. Finding videos to audit is separate: each accepted trending-videos call is $0.10, so six search calls add $0.60.
Both the analysis and caption docs describe these as fixed estimates that Sume reserves and captures per accepted job, and they point to the live catalog for the current price.
The line items
Prices are the documented fixed estimates, read 2026-10-09.
| Job | Price per accepted job | 100 videos | Duration limit |
|---|---|---|---|
| Video analysis | $0.30 | $30.00 | soft ~90 s warning, hard 300 s |
| Standalone captions | $0.20 | $20.00 | see captions page |
| Trending videos call | $0.10 | 6 calls: $0.60 | not per video |
Duration is the gate, not the price
A clip longer than about 90 seconds still succeeds but can return the warning low_confidence_long_video. A clip over 300 seconds is rejected after the probe with a stable duration error. Because the price is flat per accepted analysis, a rejected long video is not a cost but is a wasted step; filter by length before you submit.
Rejections that come after acceptance matter for the ledger. Failed jobs release their holds, so the audit total can end below $30 if some clips fail.
Order and concurrency
A hundred analyses is more than any plan's accepted capacity except Scale (120). On Pro, submit in waves of 24; use jobs_wait with include_results: true in MCP so each wave returns its results without a second read. The budget for a wave on Pro is 24 x $0.30 = $7.20 held at once.
If you want only the captions of the ten best performers, run analyses on all 100 ($30), pick ten, then caption those ten ($2.00): $32 instead of $50.
What the totals leave out
The $50 total covers the two per-video jobs and nothing else. Importing the clips so Sume can read them is a separate step, and any agent turns that orchestrate the audit are metered separately from the generation lines. If you drive the audit from an agent, read debited_usd_micros for the run or thread, which includes the LLM rows, and compare it with the $50 you computed from the fixed estimates. The difference is the orchestration cost, which is the figure to optimise if you repeat the audit weekly.
Doing it in two passes
A cheaper protocol is two passes. Pass one runs analysis on all 100 for $30. Pass two captions only the videos worth studying, for example the top 15, at 15 x $0.20 = $3.00. The total is $33 instead of $50, and the saved $17 buys 56 more analyses. Keep the trending-videos calls separate in your log, because they are per call, not per video.
Sources
Related posts
More in Pricing
- Avatar creation is $0.95: its share of 1, 2, 4, 10 and 25 clips
A new Sume avatar costs $0.95 once. Spread over 15-second standard clips at $2.76 each, it adds 24 cents per clip at 4 clips and under 4 cents at 25.
- Cancel a queued Sume video job: refund rules and the 409 after start
A queued Sume job can be canceled before it starts; after that cancel returns 409 job_generation_already_started. What is held, with a 30-second example.
- Cancel 16 of 20 queued Seedance 2.5 clips: what you keep
Twenty 10-second Seedance 2.5 clips reserve $115.56. Cancel the 16 still queued and $92.448 is released; the 4 running clips complete and cost $23.112.
- Cheapest 1080p Sume video model for 10, 12, 20 s: $1.875/$2.40/$5.00
Cheapest listed Sume video model at 1080p for 10, 12 or 20 seconds: Gemini Omni Flash 1.1, MiniMax H3 Max, then Wan 3.0, with arithmetic and a Python filter.
Written by Sume