Support

Email dashcut@dashmade.app and include your macOS version and what you were doing. Most questions are covered below.

Setting up a language model

Clip finding needs a language model. Open Settings ▸ Model, pick a provider preset, paste your API key and choose a model from the list DashCut fetches. Then press Test Model Connection. It checks the endpoint, your key, the model name and that the model supports structured (JSON schema) output, so problems show up before you import anything.

On OpenRouter, only models that support structured outputs work, and DashCut filters its list accordingly.

Using a local model (Ollama, LM Studio)

Choose the Local preset. The base URL is the API root including the version, for example http://localhost:11434/v1 for Ollama. No API key is needed. For a self-hosted server with a self-signed certificate, the Custom preset offers an option to allow untrusted TLS certificates.

Subscriptions

DashCut Pro is managed by Apple. To cancel or change plan, open your App Store subscriptions. If you have subscribed but the app still asks you to, choose Settings ▸ Subscription to refresh. Existing projects always stay viewable, even without a subscription.

Troubleshooting

A project fails after I moved or renamed the video

DashCut remembers the source video by location. Move it back, or re-import it.

Transcription says the audio is silent

The file's audio track contains only silence, which some cameras produce when no microphone was recorded. Check the file in another player.

A stage failed or the result looks wrong

Use the project's Re-run menu to repeat transcription, speaker identification or clip finding without re-importing.

Speaker labels look off

Speaker identification handles up to four speakers. Quiet or distant voices are the most likely to be missed.