Automating with AI agents
What can an AI agent actually do with my podcast?
It runs Cast for you. Cast's First Pass does the cleanup, the agent adds what you asked for on top (a target length, topics to drop, a batch of episodes), and the result opens in the Cast editor, where you finish by ear.
Ask for the outcome, not the steps: "Make a 15-minute episode from this 33-minute interview, keep the intro, put a calm bed under it." The agent imports the file, transcribes it, plays you the First Pass preview, applies First Pass to the whole episode once you say yes, cuts it down to what you asked for, builds the project and exports it.
Three layers, each doing what it is best at. First Pass is Cast's own editor: it hears the audio, so it catches every um, false start and repeat, keeps the words that carry meaning, writes chapters and scores a music bed for the episode. The agent follows your instructions: length, topics, which parts to keep, many episodes at once. The Cast editor is where you finish with full control: swap or regenerate the music, set the ducking, fix any word as text.
An agent without First Pass works from the transcript alone and cannot hear the recording, so it cuts only clear hesitations and leaves "so", "like" and "well" alone unless you name them. For the cleanest result, let it use First Pass.
Everything destructive is a proposal first. Cut planning is a dry run: you see every cut with its timestamp and the exact words being removed, and nothing touches the audio until you say build.
The agent's pass opens in Cast as normal work — the cuts land both on the timeline and as word-anchored edits in the Voice Editor's transcript. So you can let an agent do the boring first pass and then finish it by hand, editing the text like you always do.
- Cut by meaning"Remove every 'you know' and shorten pauses over a second, but do not touch the first ninety seconds."
- Clean the voiceAI noise and room removal, or isolate a voice from a noisy recording.
- Music with a licensePick from tracks you already own, search staff picks by vibe, or generate a full-length bed. Cast ducks it under speech automatically.
- Finish itChapters, and loudness for the platform you publish to. MP3 on every plan; WAV from Plus; per-lane stems on Max; the creator commercial-license PDF on Plus and Max.
Can an AI agent remove filler words and long pauses for me?
Yes. The best way is First Pass, which the agent runs for you: it hears the audio and judges every filler in context. The agent plays you the preview first and quotes the price for the whole episode before anything is charged.
Without First Pass, the agent finds hesitations and silences from the transcript and shows you every proposed cut with its timestamp and the exact words being removed. Nothing touches the audio until you say build.
You can steer it in plain language: shorten only pauses over a second, keep the first ninety seconds intact, remove every "you know" but leave "like" alone.
Can I take over after the agent has finished?
Yes — that is the design. The agent writes its pass into Cast as ordinary work: the cuts appear both on the timeline and as word-anchored edits in the Voice Editor's transcript, so you can reopen the episode and keep editing it as text exactly as if you had made the cuts yourself. Every answer the agent gives you ends with a link to that project.
Where does the music come from, and can I publish with it?
From Mubert, licensed. An agent can reuse a track you already generated, search staff picks by describing the vibe, or render a new full-length bed, and Cast ducks it under speech automatically. Exports on Plus and Max include the creator commercial-license PDF.
An agent cannot hand you the bare music file — the bed goes under your episode, and the track itself stays inside Cast.
Last updated