How to Turn Your Content into Social Posts for FREE

Every webinar you run is packed with insights, stories, and expert opinions that your LinkedIn audience would love to see. But most teams let the recording sit in a shared drive and never touch it again. The content is there, you just need a repeatable process to extract it and turn it into posts that actually perform.

In the video above, we walk through the complete process of taking a real webinar recording and turning it into five LinkedIn posts, from transcript to finished content ready to publish. This written guide follows the exact same workflow step by step, so you can follow along at your own pace.

The posts you create from this process can serve two purposes: promoting the on-demand or recording version of your webinar, or sharing standalone insights and takeaways from the session that deliver value on their own. Both approaches work well on LinkedIn.

What You'll Need (Tools Overview)

This workflow uses a combination of free and low-cost tools. The only paid element is an AI chatbot subscription (Claude, ChatGPT, or Gemini), though you can use the free tiers of any of these. A paid plan gives you better output quality, longer context windows, and access to features like Claude's "Projects."

• AssemblyAI (free tier) - to transcribe your webinar video into a text file. AssemblyAI gives you $50 worth of free credits when you sign up, which is enough for many hours of transcription. Rev.com is an alternative that offers 300 minutes free.
• Claude, ChatGPT, or Gemini - to analyze your transcript and generate LinkedIn posts. In the video, we use Claude with the Projects feature and Sonnet 4.5 model. You can achieve similar results with ChatGPT's Custom GPTs or Gemini.
• LinkedIn Post Preview tool - to preview and format your posts before publishing. We use linkedinpreview.com, a free tool that shows you exactly how your post will appear on LinkedIn, including where the "See more" button will cut off your text.
• Your webinar MP4 file - the original recording from your webinar, keynote, or virtual event.
• Example LinkedIn posts (your "style reference") - a collection of 15-20 LinkedIn posts you admire, saved in document files. These train the AI to write in a style you actually like. More on this below.

Step 1: Generate a Transcript Using AssemblyAI

The first thing you need is a text version of your webinar. Without a transcript, you are stuck rewatching the full recording every time you want to pull a quote or idea, which is exactly why most teams never get around to repurposing their webinars.

Choosing a Transcription Tool

We tested two transcription services for this workflow:

• AssemblyAI (recommended) - gives you approximately $50 worth of free credits when you sign up, which is enough for many hours of audio. This is the tool we use in the video because of the generous free tier and high accuracy.
• Rev.com - offers 300 minutes of free transcription. If your webinar is under 300 minutes (which it almost certainly is), Rev is a solid alternative.

Step-by-Step: Transcribing with AssemblyAI

Here is the exact process we followed in the video:

• 1. Go to the AssemblyAI Playground and sign up for a free account if you do not have one. Navigate to the playground or upload section.
• 2. Upload your video file by selecting it from your computer. In the video, we used an Alex Hormozi recording as the example webinar. AssemblyAI accepts both audio and video files.
• 3. Configure your settings. Set the model to "Universal", which works well for most content. Leave language detection on automatic. For the other options:
  • Key terms - not needed for this workflow.
  • Speaker diarization and speaker labels - you can toggle these on. Speaker labels are not strictly necessary for generating social posts, but they make the transcript easier to read, especially for multi-speaker webinars.
  • Auto punctuation and text formatting - leave these on. They make the transcript much cleaner.
  • Speech understanding features - turn these off. They are not needed for this workflow and will save you credits.
• 4. Click "Transcribe" and wait for the file to upload and process. This typically takes a few minutes depending on the length of your recording.

While the transcript is processing, you can move on to Step 2 and set up your AI workspace. This is exactly what we do in the video, using the processing time productively.

Exporting the Transcript

Once AssemblyAI finishes processing (you will see the full text appear on screen):

• 1. Click "Export" in the top-right corner of the interface.
• 2. Select "Export transcript text" to download the transcript as a .txt file.
• 3. Save the file to your computer with a recognizable name. You will upload this file to your AI chatbot in Step 3.

Step 2: Set Up Your AI Workspace with Reference Files

This is the step that separates generic, robotic AI output from posts that actually sound like something a real person would write. Instead of just giving the AI a transcript and asking for posts, you are going to give it a style reference, a set of rules, and specific instructions on what output you want.

In the video, we use Claude's Projects feature (available on the Professional plan). A Project in Claude lets you upload reference files and set custom instructions that persist across every conversation in that project. You can achieve the same thing in ChatGPT using Custom GPTs, or in Gemini by pasting your instructions and examples at the start of each conversation.

Building Your Style Reference (The Most Important Step)

The key to getting AI-generated LinkedIn posts that do not sound like AI is giving the model examples of posts you actually like. Here is how to build your reference library:

• 1. Scroll through your LinkedIn feed and find 15-20 posts that you genuinely like. Not just the ones that go viral (those tend to look very similar and formulaic), but posts that match the tone, structure, and style you want for your own content.
• 2. Copy each post into a document. In the video, we organized these into multiple files (LinkedIn Examples 1 through 5), with roughly 5 example posts per file. You can organize them however you like, just make sure you have enough examples for the AI to identify patterns in what "good" looks like to you.
• 3. Upload these documents to your Claude Project (or Custom GPT). In the video, you can see five files in the right-hand sidebar of the Claude Project interface. These files give the AI something concrete to reference when writing your posts.

This step is what makes the difference between generic AI content and posts that feel like yours. The AI will analyze the structure, tone, length, and formatting of your examples and mirror those patterns in its output.

Adding LinkedIn Best Practice Rules

In addition to your style examples, we also uploaded a document called "LinkedIn Post Pairing Rules" - a set of best practices for posting on LinkedIn in 2026. You can create this yourself by asking Claude or ChatGPT to research current LinkedIn best practices and compile them into a reference document. This gives the AI a frame of reference for what performs well on the platform right now, including things like optimal post length, formatting conventions, hook styles, and engagement patterns.

Setting Custom Instructions

The final piece of the setup is your custom instructions - a prompt that tells the AI how to use the uploaded files and what kind of output you want. In the Claude Project, these go in the "Instructions" panel. The instructions should tell the AI:

• To reference the example posts for tone and style
• To follow the LinkedIn best practices document
• What format and structure you want for the output (e.g., "Generate 5 LinkedIn posts with different angles")
• What kind of variety you want (e.g., different hooks, different storytelling approaches, different content angles)

In the video, the instructions produce posts with labeled angles like "Strongest hook," "Most relatable with specific personal story," and "Most provocative." This variety gives you options to choose from rather than five posts that all sound the same.

"Find examples that you actually like, not just the ones that go viral. Copy those and put them into a document, and then take that document and put it directly into your project."

Step 3: Generate LinkedIn Posts from Your Transcript

With your AI workspace set up (reference files uploaded, rules document added, custom instructions written) and your transcript downloaded from AssemblyAI, you are ready to generate your posts.

Running the Generation

• 1. Open your Claude Project (or Custom GPT / Gemini chat with your instructions pasted in).
• 2. Attach the transcript file by clicking the paperclip icon and selecting the .txt file you saved from AssemblyAI.
• 3. Type your prompt: "Please create LinkedIn posts from the attached webinar transcript."
• 4. Send the message. In the video, we use Sonnet 4.5 as the model, there is no need to use Opus for this task. We also have "Extended Thinking" enabled, which lets the model think through its approach before writing. This is optional, but it tends to produce higher-quality output if you are not in a rush.

Reviewing the Output

Claude will generate multiple LinkedIn posts, each with a different angle or approach. In our test, we received five distinct posts with labels like:

• Post 1 - Strongest hook
• Post 2 - Most relatable with specific personal story that readers can visualize
• Post 3 - Most provocative / contrarian take
• Post 4 - Tactical / actionable breakdown
• Post 5 - Promotional (driving traffic to the recording)

Each post is a different way to present the same webinar content. This variety is valuable because it lets you choose the angle that fits your current content strategy, or post multiple angles over the course of a week or two.

You do not have to use all five. Pick the ones that resonate with you, and discard or rework the rest. If you are not happy with any of the outputs, you can iterate with Claude. Ask it to "make this one more promotional," "focus on this specific story from the webinar," or "rewrite Post 3 with a stronger opening line."

Step 4: Preview and Format for LinkedIn

This step is critical and often overlooked. Raw AI output, even from a well-configured project, needs formatting and polishing before it is ready to publish. The way a post looks on LinkedIn matters just as much as what it says.

Using the LinkedIn Post Preview Tool

• 1. Choose the post you want to publish. In the video, we selected Post #2 (the story-based post with a relatable personal angle).
• 2. Copy the text from Claude.
• 3. Go to linkedinpreview.com (or a similar LinkedIn post preview tool).
• 4. Paste the text into the editor. The tool will show you a real-time preview of exactly how your post will appear on LinkedIn, including the "See more" truncation point.

Optimizing the "See More" Cutoff

This is the most important formatting detail for LinkedIn engagement. On LinkedIn, only the first 2-3 lines of your post are visible before the "See more" button. Everything below that fold is hidden until someone clicks to expand. If your hook does not land in those first few lines, most people will scroll right past your post.

In the preview tool, check where the "See more" button appears. Then adjust your line breaks and spacing so that your strongest, most intriguing opening line sits clearly above the fold. You may need to:

• Shorten the first line to make it punchier
• Add a line break after the hook to create visual separation
• Remove any filler words or setup that pushes the hook below the fold

Formatting for Readability

LinkedIn posts are read on phones, often during commutes or between meetings. Dense paragraphs get skipped. Use these formatting techniques to make your posts scannable:

• Short paragraphs - one to three sentences maximum. Break up walls of text into bite-sized chunks.
• Line breaks between paragraphs - add empty lines between each paragraph. This creates "white space" that makes the post feel less overwhelming.
• One idea per paragraph - each block of text should make a single point. If you are covering multiple ideas, give each one its own paragraph.

Making It Sound Like You

Even with great reference examples and a well-tuned AI, the final post needs your personal touch. Before publishing, read the post out loud. If any sentence sounds like something you would never actually say, rewrite it. The goal is for the post to sound like you wrote it, because ultimately you are putting your name on it.

Common things to adjust:

• Replace overly formal language with how you naturally speak
• Add a specific detail or anecdote from your own experience
• Cut any sentence that feels like filler or padding
• Adjust the call to action to match your actual goal (promote the recording, start a conversation, drive newsletter signups, etc.)

Step 5: Publish and Iterate

With your post formatted and polished, it is ready to go. You can publish directly on LinkedIn, or use a scheduling tool like Buffer (free tier) to queue it up for an optimal time.

Posting Cadence

You generated five posts from a single webinar. Do not publish them all at once. Spread them across one to two weeks to keep your feed active and give each post room to breathe. A good cadence looks like this:

• Day 1 (same day or day after the webinar) - publish the most promotional post or a "biggest takeaway" recap while the topic is fresh.
• Days 3-5 - share an insight-driven or story-based post that delivers value on its own.
• Days 7-10 - publish the contrarian or provocative post to re-spark conversation.
• Days 10-14 - share remaining posts, or hold them for the following week.

Iterating with the AI

If the five posts Claude generated do not fully match what you need, you can continue the conversation. Ask the AI to:

• "Make this post more promotional. I want to drive people to watch the recording."
• "Rewrite Post 3 to focus more on the story about [specific topic from the webinar]."
• "Give me three more variations of the hook for Post 1."
• "Write a shorter version of Post 4 that works as a tweet."

Because the AI has your transcript, style reference, and instructions, every follow-up request builds on the same context. You do not need to re-explain anything.

How to Build Your LinkedIn Examples Library

The quality of your AI-generated posts depends directly on the quality of your example library. Here is a practical approach to building one:

• Spend 20-30 minutes scrolling LinkedIn with a specific eye for posts that catch your attention. Save or screenshot every post that makes you stop scrolling.
• Look for variety in your examples. Include different post types: story-based posts, tactical "how-to" posts, opinion/hot take posts, and data-driven posts. This gives the AI a diverse toolkit to draw from.
• Do not just save viral posts. Posts that get 10,000 likes often follow a very specific (and overused) formula. Save posts that you genuinely admire, even if they only got 50 likes. Your goal is to match a voice and style, not to game the algorithm.
• Organize them into documents. In the video, we used five separate documents with roughly five examples each. You can organize however you like, the key is having enough examples (15-20 minimum) for the AI to identify patterns.
• Update your library over time. As you find new posts you like, add them to your reference files. As your own style evolves, swap out old examples for newer ones that better reflect where you are heading.

Full Workflow Summary

Here is the complete process at a glance:

• Step 1 - Transcribe: Upload your webinar MP4 to AssemblyAI. Configure with the Universal model, enable speaker labels and auto punctuation, disable speech understanding features. Export the transcript as a .txt file.
• Step 2 - Set up your AI workspace: Create a Claude Project (or Custom GPT). Upload 15-20 example LinkedIn posts you admire, a LinkedIn best practices document, and write custom instructions that tell the AI how to use the files and what output format you want.
• Step 3 - Generate posts: Attach the transcript .txt file and prompt: "Please create LinkedIn posts from the attached webinar transcript." Use Sonnet 4.5 with Extended Thinking for best results. Review the 5 posts generated.
• Step 4 - Preview and format: Copy your chosen post into linkedinpreview.com. Optimize the hook above the "See more" fold. Format with short paragraphs and line breaks. Read it out loud and adjust anything that does not sound like you.
• Step 5 - Publish: Post directly on LinkedIn or schedule with Buffer. Spread your 5 posts across 1-2 weeks. Iterate with the AI if you need different angles or adjustments.

The entire process, from uploading your video to having five finished LinkedIn posts, takes roughly 20-30 minutes once your AI workspace is set up. The initial setup (building your example library and writing instructions) takes an extra 30-60 minutes, but you only do it once. After that, every new webinar follows the same 20-minute workflow.

This process works well and produces great results. However, if you are running webinars regularly, weekly or even monthly, the time investment adds up. Setting up transcription, managing AI prompts, formatting posts, and previewing them across multiple webinars can easily consume a full afternoon each week. For teams that need to scale their content production, automating this workflow can free up hours while maintaining the same quality and brand consistency.