OpenClaw for creators and podcasters

OpenClaw (github.com/openclaw/openclaw) is a self-hosted personal AI assistant you run yourself. On its own it's a general assistant; the Content Engine Pack teaches it your repurposing workflow as six connected skills. Point it at an episode transcript and it runs the chain: transcript-clean produces speaker-labeled, timestamped text; show-notes writes a titled summary with chapters, takeaways, and linked resources; clip-finder surfaces the 5-8 most clippable moments with timestamps and a hook line for each; social-repurpose turns those into an X thread, a LinkedIn post, and shorts captions; newsletter-draft drafts an issue with a subject line and CTA; publish-checklist runs a per-episode list before you go live. One recording becomes a week of drafts.

Straight answer: it writes text and finds moments — it does not edit audio or cut video. clip-finder gives you timestamps and a reason each moment works; you (or your editor) make the actual cut in Descript, Premiere, CapCut, or with ffmpeg. Every social post, show-notes doc, and newsletter it produces is a first draft, not a finished, fact-checked publish. Treat it like a fast, tireless assistant editor: it removes the blank-page problem and the "where are the good bits" problem, then hands you something to trim. Nothing is guaranteed and nothing publishes itself — you keep the final read.

The pack operates on a transcript, so transcription is upstream of it. Most creators wire OpenClaw to Whisper (whisper.cpp or a hosted Whisper API) to turn episode audio into the raw transcript that transcript-clean then formats. To actually export the clips clip-finder points to, you'll want ffmpeg installed so OpenClaw — or you — can slice the source file at those timestamps. Neither tool ships in the pack; both are free and well-documented. The included setup guide walks the wiring, but you provide the OpenClaw install, your own model/API keys, and these two upstream tools. If you already transcribe in Descript or Riverside, paste that transcript in and skip Whisper entirely.

Fair question, and the honest answer: many individual skills are free to install à la carte from ClawHub, OpenClaw's public registry. If you enjoy assembling, prompt-tuning, and wiring skills together, do that — you don't need us. The Content Engine Pack sells the assembled result: six skills chosen to hand off to each other cleanly, tuned for the podcast-to-everything job, packaged to the OpenClaw AgentSkills spec, plus a setup guide so it runs the same day. It's the "it just works for my show" shortcut, not secret functionality. It's $39 with a 14-day refund. Clawmart is an independent storefront and is not affiliated with the OpenClaw project.

Say you drop a 60-minute interview. OpenClaw transcribes it via Whisper, then transcript-clean labels speakers and timestamps. show-notes returns a titled description with chapter markers you can paste into your host and YouTube. clip-finder flags, say, seven moments — a hot take at 12:04, a story at 28:30 — each with a hook line, so you know what to cut before you open your editor. social-repurpose spins those into a launch-day X thread, two LinkedIn posts, and shorts captions. newsletter-draft gives you an issue to edit down. publish-checklist reminds you to add links, alt text, and the episode number. You review, trim, and ship — the drafts exist instead of a blank calendar.