How to launch a podcast episode with Claude Code and ContentStudio

Promoting a podcast episode on social media is not a creative problem. It is a process problem. The creative work happened in the recording. What happens between a finished episode and a scheduled queue is a series of decisions that either happen deliberately or happen badly. This guide gives those decisions a structure: a seven-step workflow that uses Claude Code to plan the campaign and the scheduler to prepare, approve, and schedule every post.

Podcast launch workflow using Claude Code and ContentStudio

The single reason most podcast episode promotion campaigns fall apart is starting with incomplete inputs. Claude Code can only work with what you give it, and the scheduler can only publish to accounts that are connected. Both of those foundations need to be solid before you write a single word of post copy.

The Episode Packet

Assemble one document before opening a terminal. This is your source of truth. Every prompt, every draft, and every scheduled post must trace back to it.

• Episode facts: Show title, episode number, host and guest names, publication date and time, timezone, final episode URL, and an approved short description.
• Evidence: The verified transcript or detailed show notes, timestamps for any moments you plan to quote or clip, factual claims that cannot be paraphrased, and the exact spelling of names and product terms.
• Distribution constraints: Which accounts are connected in ContentStudio, the intended audience per account, brand voice guidelines, approved CTAs, the episode link with UTM tracking parameters, and written guest tagging permission.
• Assets: Cover art and edited vertical clip files, with a clear record matching each clip to its timestamp. If a captioned vertical video does not exist today, a Reel or Short is not in the plan.
• Governance: Draft owner, fact-checker, guest or legal approver if needed, final scheduler, and a hard “do not publish before” time tied to the episode going live.

Connected Accounts and CLI Setup

Install the CLI globally and add the agent skill that teaches Claude Code how to use every ContentStudio command automatically. This is a one-time setup.

⚠️  Security note:  Never put a real API key in a screenshot, a shared project file, or any document that leaves your machine. The cs_ prefix in these examples is a placeholder. Your real key lives in ContentStudio under Settings > API Keys.

Run accounts and list, and look at what comes back. If a channel you planned to use is not in that list, it is not in the campaign until it is connected through the ContentStudio dashboard. A plan built around a disconnected account will fail on publication day.

What Claude Code and ContentStudio Each Do

Before the steps, it helps to be precise about the division of labour, because this is where AI-assisted workflows get described too vaguely to be useful.

• Claude Code is the planning and analytical environment. It reads your episode packet, audits it for gaps, identifies distinct angles from the source material, structures the post data per channel, and prepares the CLI commands for you to review. It does not upload media, connect accounts, or publish anything. Every action that touches ContentStudio’s API happens only after a human reviews and approves what Claude Code prepared.
• ContentStudio’s CLI is the connected publishing system. It handles account discovery, media uploads, draft creation, approval routing, scheduling, and queue verification, all connected to the same API your ContentStudio dashboard uses. Your social media manager can review drafts in the dashboard even if they have never touched a terminal.

The handoff moment is the campaign matrix. Claude Code produces it; a human approves it; only then do CLI commands run. This separation is what makes the workflow auditable at every stage. For teams who prefer a conversational approach without the terminal, the Claude scheduling guide and the ContentStudio MCP server cover that path.

Step 1: Audit the Episode Packet Before Writing Anything

Skipping the audit is the most common mistake in podcast episode social media promotion. You jump straight to generating post copy, and seven polished captions later you discover the guest quote does not match the transcript, or the “listen now” LinkedIn post is scheduled for 7:30 AM when the episode URL does not go live until 9:00 AM. The audit catches both of those problems in five minutes.

Give Claude Code the episode packet and use this prompt:

Read the approved episode packet. Before generating any post copy,

produce an audit report flagging:

•   Missing facts: URL unconfirmed, publication time unclear, timezone not specified, UTM parameters absent
•   Quotes that cannot be matched to a specific transcript timestamp

• Assets listed but not confirmed as finished files

•  Channels in the planned campaign with no connected account

•   Guest tagging without written permission or a confirmed handle

Return only the audit report. No post copy yet.

For the customer research episode, the audit flags two blockers: the guest quote in the show notes does not exactly match the transcript at 18:42, and the guest tag is listed as “approved” in the packet but no written confirmation exists. Both need to be resolved before a single word of copy is drafted.

📌  Source-of-truth rule:  Every piece of content Claude Code generates must trace to a specific timestamp or transcript passage. If an insight cannot be pinned to the recording, it is a paraphrase that needs to be labeled as such, or it does not belong in the copy.

Step 2: Build the Angle Map

The most common mistake in podcast episode social media promotion is treating one episode as one post. The same message gets reformatted for LinkedIn, Instagram, and X. The caption shrinks, the hashtags change, and the hook stays identical. That is not a multi-channel campaign. It is one post with cosmetic variations.

A proper angle map identifies several materially different reasons a specific audience might want to listen, each sourced to a real moment in the recording, each capable of standing on its own without requiring the listener to have heard the episode first. This is the Claude social media workflow principle applied at the episode level: you open several different doors into the same conversation, not the same door painted in different colours.

Use this prompt after the audit is resolved:

Read the verified transcript and episode packet. Identify four to six

distinct angles that could each anchor a separate social post.

Each angle must:

•   Be sourced to a specific transcript passage or timestamp

•   Offer a meaningfully different reason to listen than the others

•   Be appropriate for at least one connected account

•   Stand alone for a reader who has not heard the episode

Reject any angle that restates another. Return only the angle map:

angle name, source timestamp, target audience, best channel. No copy yet.

Angle	Timestamp	Target Audience	Best Channel
Counterintuitive finding	18:42	B2B marketers who think they already do this correctly	LinkedIn host post
Practical reframe: 3 questions	24:10	Practitioners who want a framework to apply this week	LinkedIn brand or X
Guest’s own perspective	31:55	Dr. Nair’s professional network	LinkedIn guest-ready
Pre-release teaser question	07:18	Strategic leaders, marketing directors	Instagram Story
Self-contained clip moment	18:30-19:45	Broad discovery audience	Instagram Reel

Each of these is a different entry point into the same episode. The clip angle only appears here because the team confirmed a finished 75-second captioned vertical video exists for that segment. If that file were not ready, the Reel would not be in the map. The plan is built around what exists today.

Step 3: Match Posts to Channels

With the angle map approved, you assign each angle to a channel, a format, an asset, a CTA, and a go-live check. This is where the podcast episode social media promotion plan becomes operational. The governing principle is straightforward: the asset you actually have determines the format you can use.

You do not promise a Reel and then produce the clip. The Instagram scheduler and the LinkedIn scheduler can handle anything you feed them. The question is whether the asset is ready.

Post	Channel	Format	Angle	Asset	CTA	Go-Live Check
1	LinkedIn Host	Text + cover art	Counterintuitive finding	Cover art	Link in first comment	URL live? Guest tag confirmed?
2	Instagram Story	Story frame	Teaser question	Teaser artwork	Episode drops Thursday	No listen-now link: not live yet
3	Instagram Reel	Vertical video	Self-contained clip	75s clip, captioned	Link in bio	Caption, rights, account settings
4	LinkedIn Brand	Text only	Practical reframe	None needed	Episode URL + UTM	Different hook from Post 1
5	X / Twitter	Text	Teaser question	None needed	Episode URL	Character count; link placement
6	LinkedIn Guest	Text + guest clip	Guest’s perspective	Approved guest clip	Discussion question	Written guest approval; tag verified

Posts 1 and 4 are both LinkedIn but use different angles and go live at different times. Reading them back to back, they should feel like two separate reasons to care about the same episode.

If they read the same, one needs a different angle, or it is cut. Post 2 goes live before the episode does, which means no “listen now” CTA under any circumstances. The go-live check column is what enforces that. Post 6 requires written guest approval before it leaves draft status.

⚠️  When to skip a channel:  If a format requires an asset that does not exist, that format is not in the plan. The Reel appears here because a finished captioned 75-second vertical video is confirmed. If that clip falls through, the Reel is removed, and the team is not scrambling on launch morning.

Step 4: Set the Release-Week Schedule

Assigning times to posts is more than a preference. The sequence has to respect the episode publication time, each platform’s posting cadence, and the dependency between posts. No post with a live CTA fires before the episode URL is confirmed live.

Confirm the workspace timezone before scheduling anything. Run workspaces: list and check the timezone field in the output. A post scheduled at 7:30 in a workspace set to UTC fires at a very different moment if your audience is in Eastern Time.

Day and Time	Post	Platform	Rationale
Tuesday 4:00 PM ET	Instagram Story teaser (Post 2)	Instagram	Two days before release. No live CTA needed or appropriate.
Thursday 7:30 AM ET	LinkedIn host post (Post 1)	LinkedIn	30 min after episode goes live at 7:00 AM ET. URL confirmed before post fires.
Thursday 10:00 AM ET	Instagram Reel (Post 3)	Instagram	Same day, different format and audience. Not a repeat of Post 1.
Friday 9:00 AM ET	X post (Post 5)	X / Twitter	Day-after. Standalone angle that does not duplicate Thursday posts.
Friday 2:00 PM ET	LinkedIn brand post (Post 4)	LinkedIn	End-of-week practical reframe for practitioners.
Monday following	LinkedIn guest post (Post 6)	LinkedIn	Held as draft. Schedule only when written guest approval arrives.

The 30-minute gap between episode publication (7:00 AM ET) and the LinkedIn host post (7:30 AM ET) is deliberate. It gives the episode time to propagate and confirms the URL resolves correctly before the post fires.

Any “listen now” link that goes live before that check is a broken user experience. For building this into a repeatable weekly cadence, the content calendar and social media scheduling tools guide cover the broader planning structure.

Step 5: Prepare Draft Posts

With the matrix approved, you move from planning to execution. Every command in this step runs against the confirmed matrix. Nothing new is being generated at this stage.

Select Workspace and Confirm Accounts

Save the _id values for each account you plan to use. Every posts: create command needs them. If an expected account is missing from the list, connect it through the ContentStudio dashboard before continuing.

The Planning Prompt

Give Claude Code this instruction after confirming account IDs. It prepares structured post payloads for review without creating anything in ContentStudio yet.

Read the approved campaign matrix and the account IDs from the

accounts:list output. For each post in the matrix, prepare the payload:

•   Account ID and platform

• Final post copy

•  Media reference (media library ID or confirmed asset URL)

•  Destination URL with UTM parameters

•  scheduled_at in YYYY-MM-DD HH:MM:SS

•  Workspace timezone

•  Source angle and transcript timestamp

Do not create posts. Do not upload assets. Do not schedule anything.

Return the full set of payloads for review.

Always use –dry-run before posts: create

Before creating any post, preview the exact payload the ContentStudio API will receive. The MCP server and CLI both support this safeguard.

Check the account ID, the scheduled time, the timezone, and the copy. Only when the dry run output is correct do you remove –dry-run and run the live command. For the first run of any new campaign, create drafts rather than scheduled posts so your team can review the full set in the ContentStudio dashboard before anything is locked to a time.

Platform-Specific Posts Require a JSON Body

Instagram Reels, TikTok posts, and YouTube Shorts require a full JSON body file because the shortcut flags cannot specify post type or platform-specific options. This is standard ContentStudio CLI behaviour, not an edge case.

💡  One post per variant:  ContentStudio’s CLI supports posting to multiple account IDs in a single call with the -i flag. This is useful for identical cross-platform content. For a differentiated podcast campaign where LinkedIn, Instagram, and X have different hooks and assets, use separate posts: create calls so each can be reviewed and approved on its own.

Step 6: Review Drafts, Approve, and Verify the Queue

The drafts are in ContentStudio’s approval workflow. This is the human review step. It is not optional, and it is not a formality. This checkpoint is the only safeguard between a planning error and a published post.

Pre-Publish QA Checklist

Check	What to Verify
Copy accuracy	Does the copy match the approved matrix? Any typos or unintended line breaks?
Episode URL	Is the URL live and resolving right now? If not, is this post scheduled for after the confirmed publication time?
Guest tag	Is the handle exactly correct? Does written guest approval exist for this specific copy and this tag?
Media	Is the correct image or video attached? Does it match the angle and format in the matrix?
Account	Is this post going to the right account: host profile, brand page, or guest-ready?
Time and timezone	Does the scheduled_at value reflect the right moment in the workspace timezone?
CTA	Does the call to action match the post position in the sequence? No listen-now CTA before the URL is live.

When every check passes, approve and record exactly what you confirmed:

When something needs fixing:

The –comment field is the audit trail. If a post causes a problem after publication, you need to know who approved it, when, and what they confirmed. That information lives in the comment, not in a Slack message.

Verify the Scheduled Queue

After all approvals are complete, list the scheduled posts for launch week and scan the output. This is the final check that what was planned is what is actually queued.

Confirm that only the intended posts appear and that the times are correct. Confirm that no post with a live CTA is scheduled before 7:30 AM ET Thursday. A post that appears in this list and was not in the matrix is either a duplicate or a mistake. Find it and remove it.

⚠️  On AI and publishing:  Claude Code preparing a post payload and submitting it through the CLI is not a foolproof publishing gate. The approval step is where a person confirms the episode URL is live, the guest tag is verified, the copy is accurate, and the time is right. Automating past this point removes the only human safeguard in the workflow.

Step 7: Evaluate and Carry the Workflow Forward

After launch week, three separate streams of data need to be assessed separately. Social response from the platforms tells you how the posts performed as content: reach, impressions, engagement rate. It does not tell you whether anyone listened.

Episode-page traffic from your analytics tells you how many people clicked through to the episode destination. That is only measurable if UTM parameters were in place. Listening data from your podcast host tells you whether the episode found an audience. You will rarely have a clean causal line from a specific LinkedIn post to a verified listen, and that is not a reason to stop measuring.

Keep a simple per-episode record that tracks each post by channel, angle, tracked destination, reach, episode-page clicks, and listening data. Run it across several episodes and patterns will emerge: which angle consistently outperforms, which channel drives clicks relative to reach, whether the guest-ready post generates more substantive comments than the launch-day post. These observations inform the next episode’s matrix without prescribing it.

The reusable parts of this workflow are the episode packet template, the audit prompt, the planning prompt, and the QA checklist. The Claude social media tasks guide and the AI social media campaigns resource cover how to extend this kind of structured workflow across other content types. What must be rebuilt fresh every time is the content: the specific angles, the copy, the asset mapping, and the timing, because those belong to this episode, not to the workflow around it.

Conclusion

The workflow in this guide is a repeatable motion, not a one-time build. The episode packet gives you a consistent input structure. The audit catches what is missing before any copy is written. The angle map turns one recording into several distinct reasons to listen. The campaign matrix assigns every post an owner, an asset, a time, and a checkpoint. The CLI commands translate that plan into drafts a human reviews. The approval step confirms everything is actually ready. The queue verification checks that what was planned is what is scheduled.

The only thing that changes episode to episode is the content of the recording. The structure around it stays the same. By the third episode using this workflow, most teams complete the full process from packet to verified scheduled queue in under an hour. The episode is done. Now there is a workflow to make sure the right people hear it.

FAQs