How I Turn One YouTube Video into 3 Blog Posts: Castmagic Workflow 2026

The Exact Workflow — Step by Step

This runs automatically the moment a video publishes. No manual file uploads. No copy-pasting transcripts.

Step 1 — Trigger: YouTube RSS Feed

Make.com watches the YouTube RSS feed for the channel. When a new video appears, it checks the duration. Videos under 5 minutes are filtered out — transcripts from short videos produce thin drafts that require more editing than writing from scratch.

Step 2 — Send to Castmagic via Webhook

Videos over 5 minutes trigger a webhook to Castmagic’s API. Castmagic pulls the audio, generates a transcript, and runs it through its content extraction layer. The extraction prompt is configured to return three distinct blog angles — not three versions of the same post, but three genuinely different entry points into the same material.

The three angles this workflow typically produces:

• The how-to version — step-by-step walkthrough, tutorial structure
• The opinion version — one argument from the video, expanded
• The listicle version — key points reformatted as a scannable resource

Step 3 — Deliver to Notion for Review

Castmagic’s output routes to a Notion database via Make.com. Each entry includes the video title, publish date, and all three draft angles as separate fields. The database view shows which drafts are unreviewed, in progress, or published.

The draft is waiting before the video even finishes processing on YouTube’s side.

The Castmagic Extraction Prompt That Works

The default Castmagic output is a general summary. That is not what this workflow needs. The extraction prompt is configured to return structured angles, not a recap.

Copy this into Castmagic’s custom prompt field:

You are a blog content strategist. From this transcript, extract three distinct blog post angles. For each angle, provide: a working title, a 2-sentence summary of the argument, and 4-6 bullet points covering the key content. Return the output as structured JSON with keys: angle_1, angle_2, angle_3. Each angle must have a genuinely different framing — not three versions of the same summary.

The JSON output routes directly into the Notion database fields. If Castmagic returns unstructured text instead, add a Make.com JSON Parse module between the webhook response and the Notion step.

What the AI Draft Actually Looks Like

The draft that comes out of this workflow is not publish-ready. It is edit-ready — which is a different and more useful thing.

A typical Castmagic draft from a 15-minute tutorial video:

• Correct structure and logical flow — the video already had this
• Accurate terminology — pulled directly from the transcript
• Missing transitions — the spoken connectives do not translate cleanly to written prose
• Thin on examples — the visual demonstrations from the video are absent
• Generic opening — almost always needs a complete rewrite

Edit time on a typical draft: roughly 25–40 minutes. Writing from scratch on the same topic: roughly 90–120 minutes. The workflow saves the difference — not the editing work itself.

Where This Breaks

Short or unstructured videos. A 4-minute video or a casual vlog produces a transcript with no logical sequence. The AI extraction returns whatever it finds — which is usually a thin summary that requires more work than starting fresh. The 5-minute duration filter exists for this reason.

Heavily visual content. Screen recordings, design walkthroughs, and demonstration-heavy videos leave gaps in the transcript wherever the explanation was “look at this.” The draft will be structurally incomplete. Add a manual notes field to the Notion entry to flag which sections need supplemental writing.

Expecting three usable angles from every video. Some videos have one strong angle and two weak ones. The workflow always returns three — but “always three” does not mean “always three good ones.” Review all three before discarding. The weaker angles sometimes work better for social or newsletter content than for blog posts.

Castmagic API rate limits. If the channel publishes multiple videos in quick succession, the webhook queue can back up. Add a Make.com delay module (30–60 seconds) between the trigger and the Castmagic webhook if rate limit errors appear in the scenario logs.

The Reusable Checklist

Before publishing any draft from this workflow:

1. Rewrite the opening paragraph — the AI version is almost always generic
2. Add at least one concrete example not present in the transcript
3. Check that transitions between sections read as written prose, not spoken notes
4. Verify any statistics or claims — the transcript captures what was said, not whether it was accurate
5. Add internal links to related posts — the draft will not include these automatically
6. Replace the AI-generated title with one that matches your actual search target

FAQ

Does this work with Castmagic’s free plan?

The API access required for the Make.com webhook is not available on Castmagic’s free tier. The Starter plan or above is required for API integration. The workflow can be tested manually on a free account — upload the video file directly and run the custom prompt — but the automation requires API access.

Can I use a different transcription tool instead of Castmagic?

Yes. The Make.com scenario can be adapted to route to any transcription API that accepts an audio URL and returns structured text. Alternatives that work with this setup include AssemblyAI and Deepgram. The extraction prompt would need to be moved to a separate Claude or OpenAI API call in Make.com if the transcription tool does not support custom content prompts natively.

How long does the full process take from video publish to Notion entry?

Roughly 8–12 minutes depending on video length and Castmagic’s current processing queue. The Make.com scenario triggers within 1–2 minutes of the RSS feed updating. Castmagic processing adds 6–10 minutes for a 15-minute video.

What if the Castmagic output is not valid JSON?

Add a Make.com Tools → Parse JSON module after the Castmagic webhook response. If the output still fails to parse, add a fallback route that sends the raw text to a Notion field labeled “Needs Manual Extraction” and notifies via Slack. Do not let the scenario error out silently — the draft exists in Castmagic’s interface even if the automation fails.