clipfarmer

FAQ

Questions

Is this a website I sign up for?

No. clipfarmer is software you install and run on your own computer. It opens as a local web app in your browser at an address on your own machine. There is no account, no sign-up, and no clipfarmer server for anything to sign in to. This website is documentation.

Who is it for?

One person cutting clips from their own channel's material, or from material they have the right to use. It assumes a single reviewer with a consistent sense of what is funny, and it is built around that rather than around a team workflow.

What does it do with my TikTok account?

It uploads clips you approved to the one account you connected — either as a draft in your TikTok inbox that you finish in the app, or as a direct post with the caption you approved. It requests three scopes: basic account info so it can show which account is connected, and the two posting products. It cannot read your followers, comments, messages or analytics, because it has no scope that would allow it. The TikTok page has the full detail.

Where are my access tokens stored?

In a file on your own disk, alongside the rest of the job data. They are never sent to a clipfarmer server, because there is no clipfarmer server. Deleting that file ends the access immediately.

How do I disconnect it?

In TikTok, under Settings and privacy → Security and permissions → Manage app permissions. Revoking there invalidates the stored tokens straight away. Deleting the local token file works too.

Does it post automatically?

No. Every clip that goes out was approved by a person, with its caption reviewed, one clip at a time. There is no scheduler and no unattended posting. The parts that run unattended are the parts with no consequences — downloading a VOD, transcribing it, ranking candidates, rendering previews.

Does it add a watermark?

No. Nothing is added to the video except the subtitles, the layout, and — where you chose one — a header line. No logo, no branding, no end card.

Does the caption get posted without me seeing it?

No. A caption and hashtags are generated per clip as a starting point, shown to you, and editable before anything is sent. On the draft route you write or paste the caption inside TikTok itself.

What about the rights to the source video?

That is the operator's responsibility, exactly as it would be using any video editor. clipfarmer supplies no content, licenses nothing, and hosts nothing — it processes the file you gave it. The terms state this plainly.

Does it track me?

No. No analytics, no telemetry, no advertising, no crash reporting. This website sets no cookies and loads no third-party scripts or fonts.

What does it send over the network?

Downloading the VOD and chat replay from the source platform, calls to a language model, and — when you post — the calls to TikTok listed on the TikTok page. If the language model runs locally, that traffic never leaves your machine either.

How much disk does it need?

A lot, while a job is live: an 8-hour VOD plus its audio, keyframes, previews and renders runs to tens of gigabytes. Finished jobs can be archived, which releases the source video and keeps the decisions and the finished clips.

Does it need a GPU?

Not strictly. Transcription and any local model are far faster with one. ffmpeg does the rendering and will use what your machine has.

Can it post to YouTube Shorts or Instagram?

Not today. The renderer produces standard vertical video files you can upload anywhere by hand. TikTok is the only platform with a built-in posting route.

Why does a person have to approve every clip?

Because the measurements said so. Ranking quality here is scored against the clips real viewers made from the same VOD, and on that benchmark six language-model ranking variants lost to a plain statistical score. Models are good at generating candidates and bad at knowing which one is funny — so they generate, and you choose.

Is it open source?

Not currently public. If that changes it will be noted here.

Still stuck? Read the docs