ALERT FOR ABSOLUTE BEGINNERS: This guide was written with YOU in mind. Zero coding experience required. If you’ve never touched a terminal in your life — don’t worry. I’ve got you. Every step is spelled out. Just follow along.
What Is This Blog Post About?
You’ve got OpenClaw running — that magical AI lobster living inside your Telegram (or WhatsApp, or Discord). Right now, it probably chats with you, answers questions, maybe helps you write emails. That’s already pretty amazing.
But what if it could also:
- Auto-clip the best moments from a long podcast into viral-ready short videos?
- Instantly summarize any YouTube video — with timestamps — in seconds?
- Transcribe interviews or lectures so you never have to take notes again?
That’s what happens when you connect OpenClaw to WayinVideo’s API — a powerful AI video platform that does all of this programmatically. You’re essentially giving your AI assistant a video editing department.
Let’s do this.
Prerequisites: Are You Ready?
Before we dive in, let’s make sure you’ve got the basics covered.
1. OpenClaw Is Installed and Running
OpenClaw runs on a Mac (Mac Mini, MacBook, or any Mac device), a Linux server, or a Windows PC. It also needs to be connected to a messaging app — any of the following will work:
- Telegram(recommended for beginners — easiest to set up)
- Discord
- Any other IM platform that OpenClaw support
This guide uses Telegram for screenshots and demos, but everything works identically on WhatsApp or Discord. Stick with whichever one you’ve already configured.
2. An AI Provider Is Configured
OpenClaw needs to “think” using an AI model. The most recommended options are:
- Claude (by Anthropic) — exceptional reasoning, great at following instructions
- GPT models (by OpenAI) — widely used, highly capable
The setup in this guide uses GPT-4.5, OpenAI’s latest model, which handles these workflows exceptionally well.
You can also use lower-cost open-source models if budget is a concern.
Quick Check: Are You Actually Ready?
Here’s the simplest possible test:
- Open your messaging app (Telegram, WhatsApp, or Discord)
- Send your OpenClaw bot:“hi”
- If it replies —you’re ready! Move on to the next section.
- If it doesn’t reply — go fix your OpenClaw setup first, then come back.
Step 1: Get Your WayinVideo API Key
WayinVideo is an AI-powered video platform that can clip, summarize, transcribe, and reframe videos automatically. It provides an API that lets external applications — like your OpenClaw — tap into all of these capabilities programmatically.
Here’s how to get access:
1.1 — Create Your WayinVideo Account
Go to wayin.ai and sign up for an account (or log in if you already have one).
1.2 — Go to WayinVideo API Dashboard Page
Navigate to the WayinVideo API dashboard page at https://wayin.ai/wayinvideo/api-dashboard.
You can purchase WayinVideo API units directly here. Pricing comes in multiple tiers — choose based on your actual needs.
If you’re just getting started or want to experiment, $10 for 300 units is more than enough to run a few demos. Or if you’re ready to dive in, you can go straight for something like $100 for 3,500 units.
Wondering exactly how much each API call costs? No worries — it’s very transparent. The number of API units consumed for each request is included directly in the API response.
After your task finishes, you can simply ask OpenClaw: “How many units did this API call use?” and it will tell you right away. This makes it very easy to track your usage while you build.
If you’d like to explore the response fields in more detail—especially the ‘cost_usage‘ field—you can check the documentation here.
How to Generate Your API Key
Once you buy the api units:
1. Go to theAPI Keys section in your WayinVideo dashboard
2. Click“Create New Key”
4. Give it a name — for this tutorial, type:openclaw
5. Click Create API Key.
6. Copy the key immediately. It looks something like: sk_xxxxxxxxxxxxxxxx
CRITICAL WARNING: Save this key somewhere safe RIGHT NOW. A password manager, a secure note app, or a plain text file in a private folder. Keep this API key strictly private — never share it with others or expose it anywhere online, as this could lead to unauthorized use or abuse of your account.
Step 2: Tell OpenClaw About Your API Key
Now you need to give OpenClaw this key so it can use WayinVideo on your behalf.
The no-code way (recommended for beginners):
Just open your messaging app and tell OpenClaw directly. Type this into the chat:
Please configure the WayinVideo API key as a env variable in the current workspace with the value: sk-_xxxxxxxxxxxxxxxx
After setting it, please verify it is not empty and confirm to me that it’s been saved correctly.
(Replace sk_xxxxxxxxxxxxxxxx with your actual API key)
OpenClaw will handle the rest. It will store the key as an environment variable in its configuration and confirm when it’s done.
What success looks like: OpenClaw replies with something like: “Done! The WayinVideo API key has been set and is confirmed non-empty.”
The terminal way (if you’re comfortable with the command line, or if the chat method fails):
Open your terminal and run this command — replacing the placeholder with your actual key:
macOS / Linux (Recommended)
If you’re using zsh:
echo ‘export WAYINVIDEO_API_KEY=”sk_xxxxxxxxxxxxxxxx”‘ >> ~/.zshrc
source ~/.zshrc
Remember: Replace sk_xxxxxxxxxxxxxxxx with your real API key value.
If you’re using bash:
echo ‘export WAYINVIDEO_API_KEY=”sk_xxxxxxxxxxxxxxxx”‘ >> ~/.bashrc
source ~/.bashrc
Then verify it worked:
echo $WAYINVIDEO_API_KEY
If your API key prints out, congratulations — you’ve just cleared the hardest part of the whole setup.
After running this, restart OpenClaw Gateway and check that it’s working.
Run this command in your terminal:
openclaw gateway restart
Step 3: Teach OpenClaw to Use WayinVideo — No Coding Required
Here’s where the magic happens. You don’t need to write a single line of integration code. Instead, you’re going to let the AI read WayinVideo’s own documentation and figure it out.
Think of it like hiring a new employee and handing them the manual. The AI will read it, understand it, and know exactly how to call the API.
3.1 — Start With the Video Summary Feature
Let’s begin with something immediately useful: AI Video Summarization.
In your chat with OpenClaw, send this message:
Please read the WayinVideo API documentation at https://wayin.ai/api-docs/video-summarization/ and familiarize yourself with the Video Summary endpoint.
For API access, you don’t need to worry about the key — I’ve already configured it as an environment variable, so just pull it from your global environment.
Once you’ve reviewed the docs, go ahead and summarize this YouTube video for me:
[paste any YouTube URL here]
Please make sure the summary includes timestamps.
What happens next:
- OpenClaw fetches and reads the API documentation
- It understands how to make the API call
- It calls WayinVideo’s summary endpoint with your YouTube URL
- It returns a structured summary with timestamped key moments
Pro Tip: The timestamped summary is arguably the best part. Instead of scrubbing through a 2-hour podcast, you get a map of exactly where the interesting parts are. Students, researchers, content creators, and busy professionals all find this feature life-changing.
Bonus Tip: Turn Repeated Workflows into a Skill
Whenever you find yourself running the same workflow repeatedly, it’s worth locking it in as a reusable Skill in OpenClaw. That’s exactly what we’ll do here with our video summary workflow.
Send this message to get started:
Please use the Skill Creator to save the process we just walked through as a new skill called video-summary. Going forward, whenever I send you a YouTube link and ask for a summary, automatically trigger this skill and generate the summary.
Once it’s saved, you never have to re-explain the workflow again — just drop a link and go.
Now that the skill is set up, using it couldn’t be simpler. Just drop the video link and ask Claude to summarize — no extra setup needed. Try sending something like:
Please use the video-summary skill to summarize this video: https://www.youtube.com/xxxxxxx
That’s it. The skill handles the rest.
Step 4: Explore All Four WayinVideo Capabilities
WayinVideo’s API currently offers four core capabilities. Here’s what each one does:
Bonus Tip: Scale Your Skills Instantly
Since we’ve already built the video-summary skill, Openclaw now fully understands the underlying workflow pattern. That means we can use it as a template — just feed Openclaw the API docs for any other WayinVideo capability, and it can replicate the structure and spin up a new skill immediately.
Send this prompt to build all of them in one shot:
Now that we’ve learned how to work with the WayinVideo API, let’s use the video-summary skill as a template and create the following new skills. Here are their names and corresponding documentation:
- ai-clipping — https://wayin.ai/api-docs/ai-clipping/
- find-moments — https://wayin.ai/api-docs/find-moments/
- video-transcription — https://wayin.ai/api-docs/video-transcription/
Openclaw will read each doc, mirror the skill structure it already knows, and have all four skills ready to use — no manual setup required for each one.
1. Video Summary (with Timestamps)
What it does: Takes any long video — a YouTube link, a lecture recording, a webinar — and produces a structured summary with timestamped key moments.
Best for:
- Students who want to review a lecture quickly
- Researchers doing topic surveys
- Anyone who needs to understand a long video without watching all of it
Example prompt:
Use video summary skill to summarize this video and include timestamps: [YouTube URL]
2. AI Clipping: MOST RECOMMENDED
What it does: Analyzes your video and automatically identifies and extracts the most engaging, shareable, “viral-worthy” moments — then creates short clips ready for TikTok, Reels, or YouTube Shorts.
Why it’s a game-changer: Content creators who used to spend 4–8 hours manually clipping a podcast episode can now do it in seconds. The AI detects energy, emotion, and key talking points automatically.
Best for:
- Podcast creators repurposing content for social media
- Marketers extracting testimonials or highlights from long-form content
- Anyone who wants to build a short-form content library without editing skills
Example Prompt:
Use the ai-clipping skill for this video: https://www.youtube.com/watch?v=p7iwXvBnbIE&t=141s
Find the top 3 clips with the highest viral potential, export them in TikTok vertical format with AI reframe and animated captions.
Download all clips and send to me the file through gateway : MEDIA: /absolute-path-to-file
Here’s what’s actually happening under the hood when you send that:
Openclaw calls the WayinVideo API, which processes the video and scores every possible clip by viral potential. It then picks the top 3, automatically crops each one to 9:16 vertical format, uses AI reframing to keep the main subject centered throughout, applies smart auto-layout, and layers on animated captions — all in one pass.
When it’s done, the finished clips are returned directly to you.
You don’t configure anything. You don’t touch a video editor. You just send the link and collect the results.
3. Video Transcription
What it does: Converts speech in a video to accurate, editable text — complete with punctuation, timestamps, and speaker labels.
Best for:
- Turning interviews into articles
- Creating searchable records of meetings
- Building subtitle files for content localization
Example prompt:
Use video-transcription skill, transcribe this video and return the text with timestamps: https://www.youtube.com/xxxxx
4. Find Moments (Semantic Video Search)
What it does: Lets you describe a specific moment in natural language, and WayinVideo’s AI finds it in the video for you. No more scrubbing timelines.
Example query: “Find the moment where they discuss pricing strategy” — and it jumps you right there.
Best for:
- Journalists reviewing interview footage
- Analysts searching through webinar recordings
- Anyone who needs a specific moment from a long video
You have two options here: simply retrieve the timestamps for the relevant moments, or go a step further and download the actual clips directly.
Example prompt:
Use find-moments skill, in this video https://www.youtube.com/watch?v=p7iwXvBnbIE&t=141s, find all funny moments.
Download the top 3 clips and send to me these files through gateway : MEDIA: /absolute-path-to-file
Step 5: Uploading Your Own Video Files
Not all content lives on YouTube. WayinVideo also supports direct file uploads, which means OpenClaw can process:
- Videos you’ve recorded locally
- Downloaded video files
This is where the API really opens up. Any video file can be processed through the WayinVideo API — not just YouTube links. Simply ask Openclaw to upload your local file first, grab the identity value returned from the upload, then pass that value as the video_url when submitting any task.
That one small step unlocks everything: if you can upload it, you can run the full suite of WayinVideo capabilities on it.
Example prompt for local files:
I want to upload a video file to WayinVideo and then summarize it.
Read this document for uploading a video file to WayinVideo. https://wayin.ai/api-docs/upload/
Just hand the local file directly to OpenClaw. It will handle the upload to WayinVideo, fetch the summary once processing is complete, and deliver the result straight back to you.
Tips & Hacks
- Chain the features together. Start with a Summary to understand the video, then use Find Moments to locate specific segments, then clip those segments. This is a complete automated video workflow — from raw content to shareable clips — triggered by a chat message.
- Use it from your phone. OpenClaw lives in your messaging app, which means you can fire off a video summary request while commuting and have the results waiting when you arrive. No laptop needed.
- Keep your API key safe. Treat your WayinVideo API key like a password. Don’t paste it into shared documents or public channels.
- Monitor your usage. If you’re on a paid plan, check your WayinVideo dashboard periodically to understand how many API credits you’re using. Heavy batch processing can consume credits quickly.
- Test with short videos first. When you’re still learning the setup, test with videos that are 5–10 minutes long before throwing a 3-hour podcast at it.
FAQ
Q: Do I need to know how to code to do any of this?
A: No. The only “technical” thing you do is copy-paste your API key. Everything else is done by telling OpenClaw what you want in plain English.
Q: What if OpenClaw can’t find the API documentation automatically?
A: Paste the documentation URL directly into your message to OpenClaw, like: “Please read this URL and learn how to use the API: https://wayin.ai/api-docs/introduction/“
Q: Can I use this with WhatsApp or Discord instead of Telegram?
A: Absolutely. OpenClaw works the same way regardless of which messaging platform you use. The prompts and setup are identical.
Q: What AI model works best for this?
A: Claude Sonnet and GPT-4.5 models are the most reliable for reading API docs and generating correct API calls. If you’re using a lightweight open-source model, results may vary.
Q: My OpenClaw didn’t respond when I sent it the API key. What now?
A: Try the terminal method in Step 2. If that also fails, make sure OpenClaw is actually running (check your terminal or system tray) and try restarting it.
You’re All Set!
You’ve just handed your AI assistant the keys to a full video editing department — WayinVideo’s AI doing the heavy lifting, and OpenClaw orchestrating everything through a simple chat interface. No editing software, no timeline scrubbing, no export queues. Just you, a messaging window, and a very capable AI lobster. Send a YouTube link, get a timestamped summary, clip your first video, or chain all four features together and see what the pipeline can really do.
And honestly? This is just the beginning. With the right APIs connected to it, OpenClaw becomes a genuinely powerful engine for AI-driven video work — and WayinVideo is only one example of what’s possible. Go explore and see where it takes you.
