Drop in one raw, unprocessed call or meeting transcript and get back a sourced blog draft shaped to get found in search. It mines the gold first, then on first run it wires into your inbox, your keyphrase list, your blog, and your voice, and runs your version every time after.
← All Claude SkillsYour best calls are full of gold, and a raw transcript buries it. This skill mines that gold first, matches it to a real search keyphrase, and writes a sourced draft where every heading is a question people actually search. The first time you run it, it interviews you and designs itself into your world: where your transcripts land, your keyphrase list, your blog, and your voice. Every run after that, it just works. Copy the two blocks below and give them to your Claude.
Step 1. Copy the skill block below.
Step 2. Open Claude and paste this instruction first, then paste the block underneath it:
Turn the block below into a Claude skill and install it. Create the skill file exactly as written. Then personalize it with smart defaults from what you know about me: my business, my voice, my current priorities, and where I keep my work. Confirm when it is installed and tell me how to run it.
--- name: transcript-to-blog description: Turn one raw, unprocessed call or meeting transcript into one sourced, search-ready blog draft. The skill runs a Gold Mine extraction pass first, matches the gold to your keyphrase list under a Keyphrase Contract, enforces a Hardened H2 Gate so every heading is a real search question, then auto-promotes the draft to ready or holds it with named failures. On first run it interviews you to wire into your transcript inbox, your keyphrase list, your blog, and your voice. Use whenever you say "run the transcript-to-blog pipeline," "draft a blog from this transcript," "transcript to blog," or drop a raw call transcript and want a sourced post. --- # Transcript To Blog: Raw Transcript To Sourced Post Take one raw, unprocessed call or meeting transcript and turn it into one blog draft that is sourced from real lived experience, written in your voice, and shaped to get found in search. The skill mines the gold first, matches it to a validated keyphrase, writes the draft against a Keyphrase Contract, and gates every heading through the Hardened H2 test before it hands you a finished draft. One transcript in. One sourced, search-ready draft out. No source material means no draft. Source comes first, keyphrase comes second. ## First Run: Design Your Version The first time you run this skill, set it up. Ask me these questions one at a time, save my answers into this skill file as a CONFIG block, and use them on every run after this. If the CONFIG block is already saved, skip straight to the workflow. 1. Where do your raw call and meeting transcripts land? (A transcript inbox, a Notion database, a Fathom or Otter folder, a Google Drive folder, or a transcript you paste.) This is where I pull the raw source. 2. Do you keep a validated keyphrase or topic list, and where does it live? (A spreadsheet, a Notion database, a doc, or none yet.) This is the menu of search targets I match the gold against. If you have none, I propose the real search question the gold answers and flag it for your review. 3. Where does your blog live and how does it publish? (WordPress, a static site, a Notion page, a Google Doc handed to a publisher, or back to you in the chat.) This tells me the draft format and where to check for duplicate slugs. 4. Whose voice should the post be written in, and are there names or words to avoid? This keeps every draft on-brand and safe to publish. 5. Where should drafts land for review before they go live? (A drafts database, a folder, or back to you in the chat.) This is where the finished draft waits for your yes. If I ever say "redesign transcript to blog," run this interview again. ## Input - A raw transcript link or ID from the inbox you named in setup (I pull the transcript) - A pasted raw transcript from any call or meeting - "Pull my most recent transcript" (I find the latest unprocessed source from the inbox you set up) ## The Pipeline (Per Transcript) Run these stages in order. Finish the full pipeline for one transcript before you start the next. ### Stage 1: Gold Mine Extraction (runs first) Your source is raw. Nobody has processed it yet. So the first move is a Gold Mine pass over the whole transcript. Read it end to end and pull: - Key insights, the original ideas and frameworks worth stopping for - Quotable lines in the exact words spoken, these become the heart of the draft - Frameworks and mental models named in the conversation - Questions a client or guest asked that other people are also asking - Breakthrough moments where something clicked - Stories and analogies that carry a teaching - Specific tools, tactics, and recommendations **ANONYMIZATION RULE (non-negotiable):** Never use client names, company names, industries, or identifying details in the published draft. Replace them with a founder I coach, a client building their stack, someone I was working with. If industry context matters, swap it to a different industry. Real names are fine in your private library. Anonymization applies to the published post. **Source-first rule:** If the gold has no teachable through-line, no post gets written. The gold stays saved for your library. Log it and move on. A raw transcript with nothing teachable is honest to skip. ### Stage 2: Keyphrase Match And The Keyphrase Contract Match the gold against your validated keyphrase list. 1. Name the strongest one or two nuggets from the gold, the teaching moments most worth a post. 2. Cross-reference each nugget against your keyphrase list. 3. Pick the keyphrase the nugget fully answers in the keyphrase's own language. 4. Rank by priority and low competition when your list carries those fields. **Match threshold:** A real match means the gold fully answers the keyphrase in the keyphrase's own words. If the gold is adjacent and answers a different question, that is a partial match and it gets rejected. Pick the next full match, or propose the real search question the gold answers and flag it for your review. A partial match shipped is a post that serves no strategy. **Duplicate check:** Convert the proposed title to a slug and check your blog for an existing page at that slug. If one exists, pick the next keyphrase. If the slug is free, proceed. **The Keyphrase Contract (non-negotiable):** - The exact target keyphrase appears in the H1, the first 100 words, at least two H2s, and naturally through the body. - The post and the keyphrase are the same conversation. If the draft cannot honestly satisfy this, the match was wrong. Return to the top of Stage 2. - Use the keyphrase's literal vocabulary, its actual words, through the body. The transcript supplies the stories. The keyphrase supplies the words. Both stay intact. - Log the placement count with the draft. ### Stage 3: Write The Draft Source every line from the Gold Mine output. Structure: 1. **Title (H1)** carries the exact keyphrase. 2. **TLDR block** of 3 to 4 bullets. 3. **Intro (AEO intro formula)** leads with the reader's problem or question in keyphrase-rich language. The reader and the answer engine both know within two sentences exactly what question this post answers. The personal story enters in paragraph two as the evidence. Story is the proof, it is never the opening frame. 4. **Question-based H2 sections (Hardened H2 Gate, non-negotiable).** Five to seven H2s. Every H2 is a real search-form question a person would actually type into Google. **The search-form question test:** Before you call the draft done, check every H2 against one test. Would a real person searching this topic type this exact question into Google? If no, the H2 is a narrative header wearing a question mark and it gets rewritten. Good H2s (real search queries): What is an AI operating system? How do I build an AI operating system for my business? What are the core components of an AI-first business? Bad H2s (narrative headers in disguise): The Sunday That Changed My Whole Business. What Actually Clicked. How Do You Make Sure An Agent Can Run It? Each of those fails the test and gets rewritten. Every H2 answer pulls from the transcript with real quotes or paraphrased moments. The H2 is the search hook. The body is the lived experience that earns the trust. 5. **Actionable steps** the reader takes with this. 6. **Strong close** that lands the message. **Source grounding (the trust lock):** - The draft carries at least 3 direct or paraphrased moments from the transcript, anonymized when they come from client calls. - The draft references real stories and real moments from the conversation, never hypothetical examples. - If a section cannot be sourced from the transcript, flag it and write it thin. Review decides whether to cut or expand. **Internal linking (done in the draft):** - Build the current live URL map from your blog before you write. Link only to URLs confirmed live. - Write 3 to 5 in-body links to existing posts, maximum. Anchor text stays conversational. - Links live inline within paragraphs, never a related-posts list. Log every link. ### Stage 4: Polish Pass Run your polish pass over the draft in one revision. Separate spoken rhythm from written rhythm. Give every idea its own space. Trade vague intensity for specific weight and cut filler like really, actually, so much. Title-case headers, sentence-case body. Tighten the close. Run one final voice check. One pass, no iterating. ### Stage 5: Save The Draft Save the finished draft to the destination you set up in first run, with the keyphrase, proposed slug, source transcript reference, and a short check log recorded alongside it. ### Stage 6: The Auto-Promotion Gate This is the final action per draft. Run all six checks: 1. **Keyphrase Contract** exact keyphrase present in H1, first 100 words, two or more H2s, and body. Placement count logged. 2. **AEO intro formula** first paragraph leads with the reader's problem in keyphrase-rich language, the answer is clear within two sentences, the story is paragraph two as evidence. 3. **Vocabulary anchoring** the post uses the keyphrase's literal words, no synonyms or inversions. 4. **Hardened H2 Gate** every H2 passes the search-form question test and the keyphrase or a direct variant appears in two or more H2s. 5. **Internal links present** 3 to 5 in-body links, all confirmed live, logged. 6. **Voice check** no banned words, no semicolons, no em dashes, no negation openers, reads in your voice. **The promotion decision:** - If the draft passes every check, mark it Ready. - If the draft fails any check, hold it at Draft and write the specific failures into the notes, naming each failed check and why. - A draft that failed any check never gets marked Ready. No exceptions. A held draft with named reasons is an honest, successful outcome that the review queue expects. ## Hard Rules - Never write a draft without source material. No transcript means no draft. - Never skip the Gold Mine extraction pass. The source is raw, so mining is the first stage every time. - Never skip the duplicate slug check. - Never mark a draft Ready when it failed any of the six gate checks. - Never use client names or identifying details in the published post. - Never invent transcript moments. Every quote traces back to real transcript content, anonymized. ## Voice And Tone Write in the voice you set up in first run. Direct, clear, specific, action-oriented, written to the reader as you. Short paragraphs. Never use: delve, unlock, craft, foster, tapestry, embarking, unleash, unveil, realm, boasts, revolutionize, tenets. No semicolons, no em dashes, no contrastive framing, no negation openers. ## Close After saving, confirm with a short receipt: the post title, the target keyphrase, the promotion decision, and where the draft landed. Then offer the strongest pull quote or a follow-up angle if it is useful.
Claude reads the block, builds the skill file, and installs it. From then on, you just run the skill in any conversation. No folders to find, no code to write.
Every week you see what's working, build it live, and put it to work in your business.
Here's the shape of what this skill hands you when it runs. Copy it, use it, keep moving.
You don't need to hold any of this in your head. That's the skill's job. If you're curious what it's doing for you, here's the idea underneath it.
A raw transcript holds real gold and buries it under an hour of talk. The stories are true, the teaching is there, and search never sees any of it. This skill mines the gold first, then shapes it into a post built around the exact questions people type into Google. A raw call becomes a post that gets found.
The first time you run it, the skill interviews you and wires into your transcript inbox, your keyphrase list, your blog, and your voice. Every run after that just works.
Your source is raw, so stage one is a full Gold Mine pass over the whole transcript. Insights, exact quotes, frameworks, and real client questions become the source every line is written from.
The exact keyphrase lands in the H1, the first 100 words, two or more H2s, and the body. The transcript supplies the stories and the keyphrase supplies the words, and both stay intact.
The Hardened H2 Gate runs each heading through the search-form question test. Every H2 is a question a real person would type into Google, so the headings become search hooks and the spine of the post.
The Auto-Promotion Gate runs six checks and marks a draft Ready only when all six pass. Any miss holds the draft with the named reasons, so a review queue stays honest.
Your source is a raw, unprocessed transcript, so the first move is a Gold Mine pass over the whole conversation. The skill reads the transcript end to end and pulls the key insights, the exact quotable lines, the frameworks, the questions clients actually asked, the breakthrough moments, and the specific tools and tactics. That extracted gold becomes the source material for the draft, which is why every quote traces back to something real that was said. A raw transcript with no teachable through-line gets skipped and saved, so no post is ever written from thin air.
The Keyphrase Contract is the rule that keeps the post and its search target as the same conversation. Once the gold matches a validated keyphrase, the exact keyphrase has to appear in the H1, the first 100 words, at least two H2 headings, and naturally through the body. The post also uses the keyphrase's literal vocabulary, its actual words, so the transcript supplies the stories while the keyphrase supplies the language. If the draft cannot honestly satisfy the contract, the match was wrong and the skill goes back and picks a better keyphrase.
The Hardened H2 Gate makes every H2 heading a real search question a person would type into Google. Before the draft is called done, each H2 runs through one test: would a real person searching this topic type this exact question into a search box? A heading like What is an AI operating system passes. A narrative header like The Sunday That Changed My Whole Business fails and gets rewritten. The gate also requires the target keyphrase or a direct variant to show up in at least two H2s, so the headings do double duty as search hooks and as the spine of the post.