OS OpenShorts GitHub

Comparison

OpenShorts vs OpusClip, Klap and Vizard, at entry price

All three paid tools land within fifty cents of each other once you bill yearly. What separates them is the unit they count in, what the cheapest tier leaves out, and whether you can run the thing yourself.

The three options, priced

Every paid number below was read off the vendor's own pricing page in November 2026, and each plan name links to the page it came from — check it before you buy, because these change. Entry tier means the cheapest plan that removes the watermark and lets you export.

Tool Entry price What that buys Open source Self-host
OpenShorts (self-hosted) $0 + your API calls No watermark, no cap, no per-minute billing Yes, MIT Yes, Docker
OpusClip $15/mo (Starter — monthly only; Pro is $14.50/mo billed yearly) 150 credits a month, watermark removed No No
Klap $14.50/mo billed yearly ($29 month to month) 100 clips a month, no watermark No No
Vizard $14.50/mo billed yearly ($29 month to month) 7,200 credits a year, no watermark, 4K export No No

Sources: OpusClip pricing, Klap pricing, Vizard pricing — all three checked on 18 November 2026. The project's own hosted cloud is not a competitor of its repo, so it is compared on the self-hosted vs cloud page instead of in this table.

Two things are easy to miss in that table. All three paid tools land within fifty cents of each other at entry level — and all three are the same $14.50 when billed yearly. What differs is the unit they count in (credits, clips, minutes), which is why the plans feel incomparable until you do the arithmetic on your own volume. And none of them is open source, so none has a self-hosted path at any price.

What actually differs

Billing unit. OpusClip counts credits, Klap counts clips, Vizard counts credits shared between clipping and its Agent. A "clip" and a "credit" are not the same thing: Klap's 100 clips a month allows a source video up to 3 hours, while a credit-based plan charges for the length you feed in even if it yields two clips.

What the entry tier hides. Every vendor's cheapest plan is a volume limit, not a feature limit — watermark removal and export are in all of them. Where they differ is source length and resolution (Vizard's Creator allows 10-hour uploads and 4K export; Klap caps source at 3 hours), and whether API access is included: OpusClip and Klap price the API separately, Vizard includes a clipping API and an MCP connector from the Creator tier.

Who does the work. Paid tools do the processing on their GPUs and store your source files; self-hosted means the machine, the maintenance and the data are yours. That cuts both ways: no per-minute bill, and no one else's uptime to rely on.

What you give up by self-hosting. Autopilot — clipping every new YouTube upload automatically — is not part of the self-hosted stack. On CPU, clipping is 5 to 8 minutes of processing per 8 minutes of source video, so it is a background job, not an interactive tool. And you buy your own API keys, which is where the "free" gets a bill attached.

When the repo beats all three paid tools

  • You already own the GPU. An NVIDIA card brings an 8-minute source down to roughly 50 seconds. That removes the only real advantage a paid plan has on speed, and leaves a subscription buying nothing but convenience.
  • Your volume is lumpy. The paid tier costs the same in a month when you post nothing. Ten videos one week and none the next is the shape a subscription prices badly and a local install handles for free.
  • You want the code, not just the output. None of these three is open source at any tier. MIT means you can change the reframing logic, add a provider, or keep footage inside an organisation that will not send it out.

The rest of the self-hosted case — privacy, agent workflows, no watermark, no plan ceiling — holds whether or not a competitor is in the frame, so it is argued once on the self-hosted vs cloud page rather than twice.

When a paid tool wins

  • You clip every day and do not want a server. $14.50 a month is less than the electricity and attention a self-hosted stack asks for.
  • You want it done without watching. Autopilot and scheduled publishing are the only features here that the repo cannot replace, and they are the cloud's job, not a plan's.
  • You do not want to own the failure. When a render dies at 2am, a subscription means someone else is on call — this is the honest trade, and it is why both options exist.

Three questions settle it, and they are the same three the table above quietly answers: how many minutes of source video a month, whether you own an NVIDIA card, and whether clips have to publish themselves. Under an hour of source on a machine you already own, the repo is cheaper than every plan here; past a few hours a month, the minutes stop being the bottleneck and your own time is.

If you searched for the repo, you already lean towards running it: start with the four-step install, give it one real video, and decide from that output rather than from this table. If the install is what stopped you, troubleshooting covers the nine failures behind most of it.

One last check before any of this: confirm the clone is mutonby/openshorts. Several repositories carry this name with different owners, and the home page lists which one is upstream and how the others differ.