Repurpose YouTube content with ak:youtube

Use ak:youtube to extract an authorized video's content and repurpose it into one of three bounded deliverables: a blog post, an infographic pack, or a set of social drafts. The Skill routes the selected subcommand, uses VidCap.xyz when the workflow needs video metadata, captions, summaries, screenshots, or comments, and keeps the source URL attached to the result.

Choose ak:youtube for repurposing

Use ak:youtube when

• You have a YouTube URL and permission to use the source material.
• You want blog, infographic, or social output derived from that video.
• You need captions, timestamps, screenshots, or comments as evidence for a draft.
• You can review quotations, claims, attribution, brand voice, and platform fit before release.

Choose another workflow when

• You need to create a new video, script, or storyboard. Use ak:video.
• You only need a thumbnail. Use ak:youtube-thumbnail-design.
• You need general channel strategy, upload management, or live analytics. The routed surface here is content repurposing, not channel administration.
• You do not have the rights or consent needed to download, quote, transform, or redistribute the source.

Prepare the source and provider access

Before you start:

• Complete Onboarding, and confirm Marketing Kit is installed for the current runtime and scope.
• Supply a valid YouTube video URL and the desired blog, infographic, or social subcommand.
• Confirm that the video, transcript, frames, comments, music, likenesses, and other included material may be processed for your intended use.
• Make VIDCAP_API_KEY available if the VidCap-backed extraction path is used.
• Allow outbound network requests to https://vidcap.xyz/api/v1; the bundled client authenticates with the X-API-Key header.
• Ensure Python can run the bundled client with requests and python-dotenv when the workflow calls it directly.
• Choose a writable project where the generated Markdown, HTML, PNG, and JSON assets may be created.

See Runtime adapters before assuming component-level parity.

Run the Skill

Put the subcommand first, followed by the YouTube URL and any useful creative constraints. Ask for drafts only when publishing is outside the run.

/ak:youtube blog https://www.youtube.com/watch?v=VIDEO_ID Create a sourced educational draft. Preserve timestamps, mark uncertain claims, and do not publish

/ak:youtube blog https://www.youtube.com/watch?v=VIDEO_ID Create a sourced educational draft. Preserve timestamps, mark uncertain claims, and do not publish

$ak:youtube blog https://www.youtube.com/watch?v=VIDEO_ID Create a sourced educational draft. Preserve timestamps, mark uncertain claims, and do not publish

Select a routed output

The first word is the router. Operations such as caption, comments, info, summary, and screenshot belong to the bundled VidCap client used inside the workflow; they are not additional top-level ak:youtube subcommands.

Follow the repurposing stages

1. Confirm the brief. The Skill identifies the routed output, source URL, audience, voice, platforms, desired attribution, and authority boundary.
2. Extract source evidence. The workflow can request metadata, captions, summaries, screenshots, or comments through VidCap. Availability depends on the video and provider response.
3. Select usable material. It organizes key points, timestamps, quotations, statistics, frames, and audience reactions without treating them as verified claims by default.
4. Create the selected draft. The blog, infographic, or social reference shapes the artifact and its normal output path.
5. Review content and visuals. Check transcript accuracy, quote context, source attribution, brand consistency, readable exports, and claims that need independent support.
6. Stop at the approved boundary. The Skill presents files and evidence. Uploading, posting, scheduling, or changing an account remains a separate action.

Keep rights, approvals, and external effects explicit

Do not place API keys in prompts, source content, frontmatter, or generated files. Keep credentials in the runtime's approved secret mechanism.

Review outputs and evidence

A useful run should leave you with:

• The requested Markdown, HTML, PNG, or JSON files at the reported paths.
• The original YouTube URL and enough timestamps or transcript context to trace important statements.
• Clear separation between source facts, viewer comments, generated summaries, and new marketing interpretation.
• A list of missing captions, unavailable frames, uncertain quotations, or claims that require another source.
• A short report naming provider operations performed, files created, review results, and any external action still awaiting approval.

For an infographic, inspect both the browser-rendered visual and its source data. For social output, re-check current platform limits before posting. For a blog, verify direct quotes against the transcript or original video instead of trusting the generated article alone.

Troubleshoot safely

Continue with the Marketing Kit overview, or use Installing kits for discovery and lifecycle help.

Know the current limits

• The packaged router exposes only blog, infographic, and social.
• VidCap is an external dependency. Private, removed, region-restricted, age-gated, captionless, or otherwise unsupported videos may not yield the expected evidence.
• AI summaries and article conversions can omit context or introduce errors; comments are audience statements, not verified facts.
• Model selection, provider limits, availability, latency, and cost can change outside AgentKit. The Skill does not promise a particular model or result.
• Generated drafts do not guarantee search ranking, engagement, conversion, or platform acceptance.
• Runtime projection is capability-specific; browser, image-generation, and supporting Skill availability can differ even when invocation succeeds.