No matter how extensive and well documented a set of APIs, customers can find it difficult to unlock the value of product integrations or automate manual tasks without a team of technical resources spending time on development.
There is an easier path!
Today, an array of no-code or low-code tools allows us to build complex workflows in a short amount of time and with little effort by leveraging webhooks, APIs, and AI.
In this post, I’ll walk through a powerful automated video workflow we built using the Zype Platform, Qlip.ai, n8n, Make, and OpenAI. It takes a newly published long form video and produces social clips, writes copy, keeps the team informed and posts to YouTube. By automating tasks like drafting emails, clipping and posting, we eliminate manual steps and tasks are completed faster. It also scales. It might not be a daunting task to do this once a day for a single video, but what about 10, or daily for 100 videos?
Watch Video: Building Video Workflow Automation with Zype
Why Webhooks, APIs and Workflow Automation Platforms Matter
Before we dive into the steps, let’s examine the components we’ll be using.
Application Programming Interface (APIs) enables two systems to communicate with each other through a defined interface (API Endpoints) and with an expected and structured response e.g JSON, XML.
Webhooks act like instant messengers for apps and let one application immediately notify another when something important happens, such as a new video being uploaded.
Automated Workflows Platforms are built around these webhooks and ensure that the right tasks happen at the right time. They then follow a pre-determined path and evaluate each step in the workflow for execution. With minimal coding, or sometimes none at all, teams can orchestrate complex flows using orchestration tools that connect multiple APIs without having to develop extensive code. If you’re used to having integrations coded by hand, you know it can be very time consuming. In this tutorial, I’ll be using both n8n.io and Make.
Qlip.ai (a Zype-integrated partner) specializes in automatically generating impactful social-friendly video clips in multiple aspect ratios often zeroing in on a person’s face or action in the frame. This solves a major pain point of having to manually edit or resize content for different social platforms and styled captions can be auto-generated and overlaid onto the video.
Examining the Workflow Automation, Step-by-Step
Let’s break down the actual workflow we set up:
- Video uploaded to Zype: The journey begins the moment a new video is uploaded to the Zype Platform, which handles hosting and content management.
- Webhook notification from Zype: Once the new video processes, Zype triggers a webhook to our custom web service, alerting us that there’s fresh content ready.
- Custom web service receives notification: We developed a small web service (with some help from AI) that listens for Zype’s notification of a new video. This service extracts metadata like title, description, or specific “social clip” flag.
- Check metadata and post to n8n: If the new video is flagged for social clip creation, it sends a notification to n8n, kicking off our automated workflow.
- OpenAI (GPT-4) drafts email for subscribers
n8n, upon receiving the video data, executes the first automation step and uses the metadata to draft an email notifying interested subscribers that a new piece of content is available. Note: This email could either be ready-to-send or queued for review. The goal being writing this email foundation, saving your marketing team’s time. - Slack notification to the team: The next step pings a Channel in Slack to let the team know that a video flagged for social content is now in process. This transparency ensures everyone is aware without duplicating effort and the workflow can be monitored.
- Send video file and metadata to Qlip.ai: We then send relevant video metadata and upload an MP4 file to Qlip.ai along with a callback URL for when processing is completed. Check out Qlip.ai here: https://www.qlip.ai/#features-api
- Qlip.ai callback to n8n: Once Qlip.ai finishes processing, it calls a second webhook in n8n posting a short video clips based from the original video.
- Leverage OpenAI for social post content: With the callback data, the workflow can retrieve the new clip’s URL, title, and transcription. We feed these details into OpenAI to generate an on-brand social media post title and description.
- Slack notification Update: Another Slack message fires off to inform the team that the second phase of processing (clip generation and social copy creation) is in motion. This ensures everyone knows the status of the project, plus it’s kind of fun getting alerts from your automated workflow.
- Pass content to Make.com: Here’s where we split the workflow: for final posting and authentication, we hand off the video clips and social copy to a Make.com webhook. While n8n could post directly, Make.com has streamlined integrations with certain social platforms and might simplify authentication or scheduling. Note: You may be asking why am I using two different Workflow Automation platforms together? Although n8n and Make are similar, both differ in capability and learning curve. The social nodes on n8n were more complex to configure and Make allowed me to just authenticate using my social login. Plus I wanted to get some exposure to using each for the first time.
- Make.com receives data: Make webhook is triggered, receiving the clip(s), transcription, and AI-generated social post text.
- Publish on YouTube: For our demonstration, we push the newly created clip to YouTube, setting it as Private. That way, the marketing team can review and make any final tweaks or changes. A little oversight ;)
- Ready for review: The content is now staged on YouTube, awaiting the team’s final review. If everything looks good, flip the video to Public and start tracking performance.
Now, I’d call myself slightly technical, but not a developer. It probably took me around 5 hours to configure this complete workflow and that was mainly because it was the first time I’ve used n8n and Make. Having even the limited knowledge I now have of these platforms, I’d estimate it would now take less than an hour to replicate.
Opportunities to Improve this Video Workflow
No workflow is ever perfect out of the box. Here’s some thoughts on how we might improve or build upon the current iteration:
More Robust Data Validation
Improved error checking and better verification of written social content could improve the confidence to allow the workflow to operate autonomously without review. The use of Retrieval-Augmented Generation (RAG) can help better optimize the output through referencing a knowledge base outside of its own data sources before generating a response.
Deeper Subscriber Segmentation
Instead of emailing every user on the list, we could integrate with Zype CRM or a subscriber analytics tool to pinpoint the audience that’s most likely to engage. If a subscriber watched another video in the same series, they’d be top priority for the new episode.
Advanced AI-Optimized Social Posts
Imagine drafting multiple post variations for Twitter, LinkedIn, or Instagram, each with platform-specific constraints (like character limits) and best practices. AI can tailor the format (hashtag usage, mention style, link preview) automatically.
Multiple Clips Formats
Qlip.ai (and other clipping tools) can output landscape, square, or vertical content. Qlip.ai will resize horizontal videos to vertical or square with auto-focus tracking the point of action. We can expand our workflow to push each format to the relevant channel—vertical for Instagram Reels and TikTok, square for Facebook, full HD for YouTube.
Review and Approval Steps
For organizations needing compliance checks or brand guidelines, a Slack-based approval process could be introduced. The video or social post won’t go live until an approver selects “Approved” in Slack, which triggers the final push.
Automated Scheduling
Multiple clips can be produced and the workflow can schedule them to publish at peak view times over a period of time for a constant trickle of content.
Final Thoughts
Today’s marketing landscape requires the distribution of content across numerous platforms, and for certain segments like news, at the quickest speed possible. By linking powerful APIs, leveraging AI for repetitive tasks, and integrating no-code workflow tools like n8n and Make.com, we free up time and delegate the busywork to workflow automation.
Imagine if your next big release was automatically clipped into multiple social media teasers, emailed to just the right segment of your audience, and posted to your major channels—without any intervention. That’s the vision of well-orchestrated, automated video workflows.
Let me know if you found this post interesting. If there is enough interest, we’ll make it a series and publish the workflow automation templates for review or use.