Back to Skills

podcast

Use when the user asks for podcast or a task matching the examples below. Two-host podcast video for any URL or free-form topic — 1 minute, 4 acts × ~15s, native multi-shot dialogue, optional voice cloning for Host A. Use when the user asks to "make a podcast", "podcast about [thing]", "podcast review of [url]", "two-host explainer", "interview-style clip", "two people talking on camera", "I/me and X talk about Y", or "interview with [persona] about [topic]". Native audio is the deliverable; cap

38stars7forksUpdated 8/17/2026

Security Assessment

Low Risk(89/100)

Detected risks:

Synthetic Media / Likeness(The skill can build a two-host talking video around a named real public figure, which is a synthetic-media persona use case., It mitigates this by refusing to auto-generate a photographic likeness of a real person and defaulting to a generic archetype portrait unless the user supplies an image.)
Security Score89/100

About podcast

Podcast is a Pika Labs plugin skill that produces a one-minute two-host podcast-style video from any URL or free-form topic. It is structured as four roughly 15-second acts with native multi-shot dialogue, Host A always on the left and Host B on the right, and supports optional voice cloning for Host A. Native audio is the deliverable, and captions are skipped by default because podcast dialogue tends to mistranscribe domain terms. It solves the problem of turning a link or topic brief into a conversational, on-camera explainer clip.

Parameters cover the input (URL or topic), studio background image, host portraits, host voice IDs (with Kling preset defaults), and aspect ratio; missing backgrounds and host portraits are auto-generated silently. A cost-transparency gate (estimated 6,000-9,000 credits / ~$60-$90 for four Kling v3-omni pro acts, optional image generation, a corrective retry, concat, and QA) is the only yes/no gate. The skill fires fast after that, with a pre-generation wall-clock guard and per-stage progress checkpoints, and long-running tasks are polled via task_status until terminal. A notable guardrail: when the user names a real public figure (for example in a 'me and X talk about Y' brief), the skill does not auto-generate a photographic likeness and instead uses an archetype portrait, and it documents that inline-pasted images on Claude Desktop must be provided as a hosted URL or uploaded file.

Target users are content creators, marketers, and educators who want a quick two-host explainer or interview-style clip for a topic, product, or article.

FAQ

What input does it take?

Either a URL to review or a free-form topic/brief (for example a two-person discussion prompt). Backgrounds, host portraits, and voices default to auto-generated assets or Kling presets if you don't supply them.

How much does a run cost?

The cost gate estimates about 6,000-9,000 credits (~$60-$90) for four Kling v3-omni pro 15s acts plus optional image generation, one corrective retry, concat, and QA. This is the only confirmation gate.

Can it depict a real public figure I name?

It will build a two-host clip around the persona, but it does not auto-generate a photographic likeness of a real person; it uses an archetype portrait instead unless you supply a real image.

Why are captions off by default?

Podcast dialogue mistranscribes domain-specific terms, so captions are skipped by default; native audio is the intended deliverable.

Does it support voice cloning?

Optional voice cloning is supported for Host A, and both hosts accept Kling preset or cloned voice IDs; otherwise built-in preset voices are used silently.

Install podcast

Download and extract the skill files to your .claude/skills/ directory.

Quick Setup:

  1. Copy the skill folder to .claude/skills/
  2. Claude will automatically detect and use the skill