Recently, I needed to breathe new life into a couple of old marketing videos we’d produced under the Everactive banner. The old visuals could work as a storyboard — but the branding needed a refresh and so did the voiceovers. To script new voice talent properly, I first needed accurate transcripts of the old narration.

I figured: Easy — AI can handle that.

🔍 The Problem

My first stop was the obvious one: I tried uploading the videos to ChatGPT and Gemini — ChatGPT doesn’t accept video uploads. I already had the files in Google Drive, but Gemini balked at the task. No quick wins there despite me being on paid plans for both.

I didn’t want to waste time manually transcribing two long videos line by line. So I turned my attention to OpenAI’s Whisper, a free, open-source speech recognition model that’s quite good — and runs locally on your machine.

But here’s where the next challenge popped up: installing Whisper on a Mac can be a little finicky, especially if you install Python the “normal” Homebrew way. I learned that the hard way.

⚙️ Why It Failed at First

When you install Python with Homebrew, you get the latest version (in my case, Python 3.13). The catch? Whisper and some of its dependencies aren’t ready for it yet — so pip install openai-whisper explodes with cryptic errors about wheels and build dependencies. This is painful if you’re not comfortable with Python!

✅ The Solution: Use Pyenv & an Older Python

The fix is simple once you know it:
• Use Pyenv to manage Python versions.
• Pick a stable version (I chose 3.10, 3.9 would also work)
• Install Whisper inside a virtual environment.

Here’s exactly how I got it working, step by step.

Pre-requisite, install HomeBrew for Mac.

🎯 Step 1: Install Pyenv

brew install pyenv

# Add pyenv to your shell config (Zsh)
echo 'export PYENV_ROOT="$HOME/.pyenv"' >> ~/.zprofile
echo 'export PATH="$PYENV_ROOT/bin:$PATH"' >> ~/.zprofile
echo 'eval "$(pyenv init --path)"' >> ~/.zprofile
echo 'eval "$(pyenv init -)"' >> ~/.zprofile
source ~/.zprofile

🔑 Step 2: Install Python 3.10

pyenv install 3.10.13
pyenv global 3.10.13
python3 --version  # Should say Python 3.10.13

📺 Step 3: Install FFmpeg (if you haven’t already)

brew install ffmpeg

🖥️ Step 4: Create & Activate a Virtual Environment

python3 -m venv whisper-env
source whisper-env/bin/activate
pip install --upgrade pip

🤫 Step 5: Install Whisper

pip install -U openai-whisper

🎉 Now Run It

whisper "Everactive MHM Video 1080p h264.mp4" --model base

It outputs a full text file, including timecode, you can drop right into your new storyboard or hand off to your voice talent.

📌 Takeaway

If you’re refreshing old video content, automatic transcription can save you hours — but only if your tools cooperate. Whisper is fantastic, but on a Mac, it pays to know that the latest Python isn’t always your friend.

Next time, I’ll skip the dead ends and jump straight to pyenv. Hope this saves you the same headache!

#MarketingOps #OpenAI #Whisper #macOS #ContentCreation