Spent about a day following this tutorial step by step. It took a while for me as I'm new to this but I got to the end and got a perfectly functioning automated video creation system up and running. A few thoughts for anyone who's serious about doing this: expect to spend an initial $100 or so for all the APIs. As of September 2024, I paid $50 for 120 minutes of json 2 video (i think it was their cheapest option), and ran out of make's free runs pretty quickly so had to buy a subscription for that as well. As soon as I switched to leonardo's animation mode, it ate up those points really fast - about $5 per video of about 200 words, it was generating 2-3 clips per sentence - still trying to figure out how to reduce the amount of clips generated for each story. Overall I'm really impressed with this new way of content creation. Anyone who wants to generate lots of content rapidly and doesn't mind the inconsistency of AI generated art can certainly benefit from doing this. Stephen, thanks for putting together such a great tutorial and sharing all this knowledge for free.
@@socloseagency I think he wants to find the right prompting so he can have a good result with his images with one try. This is a nice video, but it could have been better with a disclaimer that it costs a lot of money (> 100 dollar) per month to make multiple videos with this methods. You need a subscription with Airtable to run the scripts, with openAI for the API, with Leonardo to AI the images, with Eleven Labs for the audio etc etc.
@HaydenLee you can edit the Story Types prompt, where it says to generate an array of 3 strings. If you make that 1 string, you will get 1 image per scene.
Introduction and Overview (00:00:00) Setting Up the Airtable Base (00:00:56) Building the Story and Scene Tables (00:04:26) Generating Story Types and Prompts (00:06:08) Linking Stories to Scenes (00:09:04) Creating the Video API Table (00:11:29) Generating AI Image Prompts (00:16:24) Generating AI Images and Videos (00:21:54) Building Scene Audio and Video (00:47:22) Merging Audio and Video (00:58:12) Building the Final Video (01:12:22) Testing and Automation Tweaks (01:24:09) Using Animated Videos Instead of Stills (01:31:12) Final Testing and Conclusion (01:37:10)
You’ve been elevated and promoted to ‘’GENERAL’’ you’ve outgrown your previous position 🎉 CONGRATS AND THX , CARRY ON AND CONTINUED SUCCESS , AN ABUNDANCE OF SUPPORT IS DEFINITELY AFOOT.
Here's a detailed breakdown from the transcript of the steps to build the automated faceless video creation system: Set up the Airtable base: a. Create tables: Stories: Main table to manage overall video projects Scenes: Breakdown of each story into individual scenes Story Types: Define different types of stories and associated prompts Video API: Track media generation requests and results b. Set up fields in Stories table: Story ID (Auto Number) Story Type (Link to Story Types) Scenes (Link to Scenes) Source (Long text) Generated Story (Long text) Ready to Go (Checkbox) Done (Checkbox) Video (Attachment) Number of Scenes (Formula) Completed Scenes (Formula) Ready (Formula) c. Set up fields in Scenes table: Scene ID (Auto Number) Story (Link to Stories) Scene (Long text) AI Image Prompts (Long text) Videos (Lookup from Video API) Audio (Attachment) Combined Video (Attachment) Merged Audio and Video (Attachment) Number of Videos (Count) Completed Videos (Count) Ready (Formula) d. Set up fields in Story Types table: Story Type Story Prompt Text (Long text) Scenes JSON Prompt (Long text) Output Width (Number) Output Height (Number) Image Width (Number) Image Height (Number) e. Set up fields in Video API table: Job ID (Text) Video (Attachment) Scene (Link to Scenes) f. Create views for different stages of the process: Generate Story Generate Scene Videos Build Scene Audio and Video Merge Audio and Video Build Full Video Configure Make.com automations: a. Set up webhooks: Leonardo AI image generation callback Leonardo AI video generation callback Json-to-video scene completion callback Json-to-video full video completion callback b. Create main automation flows: Generate Story and Scenes Generate Scene Videos Build Scene Audio and Video Merge Audio and Video Build Full Video Implement the story generation process: a. Trigger from Airtable when "Ready to Go" is checked b. Retrieve story details from Airtable c. Use GPT API to generate full story based on source and story type prompt d. Update Airtable with generated story e. Use GPT API to generate scene breakdowns and AI image prompts f. Create records in Scenes table for each generated scene Set up image/video generation: a. Trigger for each scene without associated media b. Use Leonardo AI API to generate still images or animated videos: For still images: Make one API call For animated videos: Make two API calls (still image, then animation) c. Update Video API table with job IDs and media URLs d. Link generated media to corresponding scenes Implement audio generation: a. Use 11Labs API to generate voiceovers for each scene b. Save generated audio to Google Drive c. Create shareable link for the audio file d. Update Scenes table with audio file URL Create scene videos: a. Use Json-to-video API to combine images/videos with audio for each scene b. Configure video parameters (FPS, dimensions, zoom effects) c. Set up webhook for completion notification d. Update Scenes table with combined scene video URL Assemble final video: a. Trigger when all scenes for a story are ready b. Retrieve all scene videos for the story c. Use Json-to-video API to combine all scene videos d. Add captions to the final video e. Set up webhook for completion notification f. Update Stories table with final video URL Set up error handling and retries: a. Implement "Break" modules in Make.com for rate limit handling b. Add error filters for specific error codes (e.g., 429 for rate limits) c. Configure retry attempts and intervals Create and configure Airtable views: a. Set up filters to manage workflow stages b. Create views to trigger different automation stages c. Design views for easy monitoring and management of the process Test the entire process: a. Run test cases with different story types and sources b. Monitor each stage of the automation c. Verify output quality and consistency d. Identify and fix any issues or bottlenecks Optimize and refine: a. Adjust prompts and parameters for better output quality b. Fine-tune automation timings and error handling c. Optimize Airtable base structure and formulas if needed Optional: Modify for animated videos a. Update Leonardo AI API calls to generate animated videos b. Modify Json-to-video configuration to handle animated inputs c. Adjust timing and transitions in final video assembly Documentation and maintenance: a. Document the system architecture and workflow b. Create user guides for operating the system c. Set up monitoring for API usage and quotas d. Plan for regular updates and maintenance of the system This detailed breakdown provides a comprehensive guide to building and implementing the automated faceless video creation system, allowing for customization and expansion as needed.
Btw I did everything on video but there are also additional costs. - Airtable 24usd - Leonardo ai 9usd - eleven labs 5usd - json2video 50usd (if you dont want have a watermarks on videos)
What I learned from this video: - Automating the creation of faceless videos with animated scenes, voiceovers, and captions. - Using separate tables in Airtable to manage stories, scenes, and image prompts. - Defining story types to generate different outputs, like comic news or photorealistic images. - Triggering automation to generate and assemble stories and scenes. - Creating image prompts and animated videos for each scene part. - Generating audio and assembling it with video parts into a final product. - Configuring Leonardo API for image generation with webhook callbacks for updates. - Using webhooks to link Json to Video’s response to Airtable. Thank you Stephen this is some great content, keep going!
Stephen, I follow you from Bolivia (South America) and I love your video, I have been waiting for it for a long time, for me this is pure gold and thank you for giving us so much FREE quality content. Que Dios te bendiga
This is absolutely the first time that I blindly subscribed to a bunch of tools just to see a project through, you are the man Stephen. You should definitely do this more often, this channel will blow up for sure.
00:01 Automate building faceless videos with Make and Airtable 01:54 Automated process to create videos from stories 05:28 Setting up database fields for storytelling automation 07:17 Setting up image and video dimensions 11:12 Creating a table to store images and videos generated using the Leonardo API 12:56 Creating and managing video-related fields in the database 17:00 Automations and scene assembly process 18:52 Adding new columns and creating formulas in a storage table 22:21 Creating an automated process for generating AI-generated videos. 24:01 Ensure specific rows enter the view 27:19 Automated process to generate stories using record data and ChatGPT 29:04 Updating permissions and executing tasks with FACELESS 32:27 Creating a new row in the scenes table linked to the story for each scene generated from ChatGPT. 34:03 Automated process for generating AI images and videos 37:36 Automating video generation process with script and automation testing. 39:05 Setting up Leonardo module for generating images 42:30 Automating video creation and data entry in Airtable 44:09 Updating records with video and image URLs 47:53 Setting up web hooks and automating scene audio and video generation 49:31 Creating and using an iterator to loop through images and create JSON 52:59 Setting up video structure with necessary IDs and parameters. 54:48 Creating a web hook in Airtable for Json endpoints 58:12 Automating speech synthesis with 11 Labs 1:00:00 Automated process of updating audio and merging audio with video 1:03:40 Guide on creating scenes using JSON for video and audio elements 1:05:34 Using web hooks to automate processes with Airtable and video rendering 1:08:52 Troubleshooting and optimizing automated scene generation process 1:10:23 Automated process creating video objects and scenes with delays for successful completion 1:13:48 Automate video creation process with FACELESS Videos 1:15:34 Processing and parsing Json elements for output 1:18:59 Customizing video elements using specified settings 1:20:33 Updating the main story and testing the trigger 1:24:08 Transitioning from still images to animated videos using Leonardo 1:25:53 Setting up automation for processing video content 1:29:05 Troubleshooting the automation process 1:30:40 Automated video generation and data integration 1:34:05 Transitioning from static images to automated videos 1:35:40 Automated generation of scenes, images, and videos 1:39:09 Automated system for story types and image prompts
This is such an impressive workflow! The way you designed each table to break down scenes and different story types is brilliant. I especially loved the quote, 'It just gives me a lot of flexibility to change the output of the video and the style without having to go into the automations and tweak all of these individual nodes.' Have you considered including an automated caption styling option to add even more customization possibilities? And if you're looking to enhance your visuals quickly without the hassle, Aux Machina might be the perfect companion tool rather than Leonardo. It’s a super easy, intuitive way to generate realistic photos, so you can focus on creative freedom without getting bogged down in complex settings.
Damn what did I just watch ?????? This is mind blowing thanks a bunch Stephen. I enjoyed every little bit. First time in my dev journey I subscribed for tools just to get a tutorial done. Thanks
I just watched this video in its entirety, and it is definitely the best video of the year. Thank you so much, Stephen G. Pope, for everything! Greetings from Brazil.
Wow, i am going to be a member of your group monday. 5 week vacation going to be spent on learning this + more. This just opened my mind and ideas are flowing in. Excited, see you there!
Brilliant video. Something that would be really useful for marketers would be the ability to create these videos quickly but somehow pull footage and images from a repository of “approved” on-brand images / videos rather than using generative assets from Leonardo, stable diffusion, etc. doesn’t seem like it would be that hard to do but I’m guessing there would need to be a few alternative steps in make to search a DAM or storage location for assets with tags that match the script.
hi @EdwardAustin, i read that you have finish the tutoriel so i would like to ask you if you can help me because i cant do more after 1:06:42. When i start the automation with my last module and other side i click on test the action (airtable), they says : The scenario run was completed. but nothing its happen ( i dont have notification on my HTTP module). its hard to explain because im beginner but maybe you could help. Thank you
Hi, I'm stuck on the video and audio merger. I followed your steps and used your prompt, which generated 9 scenes. For the first scene, I got 3 images and created a 15-second video. However, when merging audio and video, I used -2 for video and -1 for audio, resulting in only two images being shown. I saw a comment suggesting to use -1 for the video, which gave me a full 15-second video, but the audio cuts off at 7 or 7.5 seconds. What am I doing wrong? Can you help?
Thank you! With this video, I am truly amazed. I expect that anyone with automation skills has found an online treasure trove. You can sell your skills or use them to earn money. Thanks again-this is a game changer!
I've watched this 3 times. It's a big confidence booster.👍💪 Just need to come up with a syllabus on learning. Checking out your skool community to start the journey. 👍💪🤯
@@StephenGPope Every time I re-watch this it makes more sense. Having a bit of trouble connecting my domain to Google business. Doesn't want to verify the dns text file. I think it's because blue host didn't have a 4 hr option. Once that's done I can get "Make" and airtable. Wow 😮 I'm realizing: I don't know what I don't know... 😕
This is sooo good... thanks heaps Stephen. It took a bit to sort out the google drive connection, but a short video on youtube sorted that out. Love all your content... so much to learn! 😊
I love this. I just joined the community. I wanted to share that if you have timestamps on your youtube videos that would be a huge help to going back and forth to specific tasks/parts
waww hands up bro ( honestly people have to pay for this ) I just finished the video and managed to create my first automation video (I set the speed to 0.5) the cost will be about 150 dollars for 200 minutes (very very cheap) i will now try to adjust the tiktok video and find a way to add a lmusci background thanks again Stephen with all my respect for you and your work
@@StephenGPope I’m having trouble figuring out how to automatically upload the final video to Google Drive after it’s generated . Could you please explain how to do this?
When you start on the second part of the automation in make , when i created the airtable get record and selected scenes and want to add a filter the action keeps pointing out to generateStory instead of generateScenesVideos - Why is that so??
Hi Stephen, No words and mind-blowing tutorial. Thanks a lot for taking the time and explaining step-by-step. The most useful and highest quality video
Re: the JSON2VIDEO errors, I was getting "Error: Movie duration cannot be zero" on the generateSceneAudioVideo action until I realized I'd missed checking the "elements" box on the Array Aggregator at 52:31. JSON2VIDEO thought my video was zero seconds because it wasn't receiving the necessary info. Just one more thing to check when troubleshooting this issue.
This was an awesome video, thank you for the content. Question, is this setup inside of the program, or would have to watch the video and go step by step again in order to implement it?
Thank you so much. That was awesome. I can’t believe how fast you talked. It’s not a bad thing I can slow you down. And watch it again, but I can’t imagine go yourself and drink of water 😂 like an hour you were like an auctioneer.😂you might be really really good at editing and I didn’t catch it, but you didn’t mess up at all like a perfect marathon sprint checked out your school that’s cool. I’ll hang out there but can’t afford it yet.
What kind of automation do you teach in your community? What applications of AI? can you give examples? I can see you're producing a short film here, so I'd like to know how this info is applicable to other businesses. Thanks!
Hi Stephen, thanks for this! I am wondering why not use Dalle 3 or Midjourney (routed through make) instead of Leonardo for the image generation part? Is it because of some price or quality reason? Then you could always send it to Leonardo later for the "movement" part. Appreciate your thoughts!
i keep getting this error - in the merge audio + video - this error comes from JSON2Video Error: Scene #1, element #1: The URL provided is not a video file and cannot be used as a source for a video element. The URL provided is an image file (codec: mjpeg). anyone knows why so ?
Cool, thanks for sharing, when you say reply about cost 100$/month for making many videos? what's the average for "many" videos? just to get some numbers. cheers!
I really don't have exact numbers for you sorry. you'll need to estimate it by going to the sites and looking at what it will take, and how long the videos are, etc. Not a simple problem to calculate.
I was wondering the same thing. I think if you are going to go large scale, this would proabably be well worth the $$. I'm going to try and figure out a $/min cost.
God video, I have a question. It seems like you already know what to do because you've made it work and are now creating the video. But I'm curious, what does your process look like when you're actually figuring things out?
@@StephenGPope I finished my first one yesterday. Having trouble with the prompt for the story not updating to the new source but I’m confident the answer is at the beginning lol. Thanks so much for your effort.!!
Curious how many operations it takes to make a video, how much the video costs in terms of eleven labs, LLM tokens and the monetization strategy for the video. I saw a SAAS company working on something like this but generating marketing videos for B2B clients
@stephengpope How much would you approximate it would cost with your automations for a 2 min video? Also, do I need to know how to code to put this entire flow together?
If the scenes generated describe a main character. How would you go about ensuring the characters visuals are kept consistent across all scenes. I have randomly generated the scenes from the story and got a slightly different result, which then describe a main character, and when those prompts are passed to leonardo, the main-character doesn't match visually from scene to scene. I think this issue is present in many story based generation. How would you make leonardo aware to keep the entire context consistent.
try repeating the character description on every prompt with detail. try sending the first photo of the character do Dalle and asking it to describe the character, that way you can use its description as the promp to the others.
much appreciated for your informative video. I have two questions: Can we change it the language to something else? Can we make the video doesn't exceed one minute?
two questions: 1. what if I dont like the narration? does it allow me to edit it before it goes through the wholr process? 2. can I generate more images per secene, or shorten the 15 seconds for each secne? the tempo right now is too slow. thanks.
Would you mind breaking down how much rev you're generating from content vs. info products? We don't need actual numbers just percentages I think would be the most useful.
Hello @stephengpope it is working perfectly! Can you explain in detail how to add music on the last step ? I tried different ways but i cant make it work in the same step, only by adding new step but loosing json2video quota. Love you 👏😎😜
So how this works? for every type of video you want to create, you have to do the whole propcess from 0? and how long videos you create like this could be?
Your work is very good, I just have one question, my video is longer than 60 seconds, which generates a json plan error, how can I limit my videos so that the duration is a maximum of 60 seconds?
Thanks for the detailed guide. I got stuck at the BUILD FULL VIDEO stage at minute 1:16:00 I run the script successfully but in MAKE it doesn't show that it produces the videos in JSON and therefore in the array aggregator there is no option to select ELEMENTS. Do you have any idea that can help me with the problem?
probably better to join my free or paid community to post the question with a screenshot so I can see what you're talking about. Debugging this via youTube comments is HARD.
Anyone know why at 35:44, at “action” it says “generateSceneVideos” but for me it still says “generateStory” this is also the same when making a filter on that module?
@@StephenGPope but the generateSceneVideos action shows up on your airtable module on make.com before you even create the automation on airtable? am i misunderstanding?
Absolutely amazing! Thanks a lot for all this value! im glad that i found you, i was literally looking for some real life example like this!, may i ask some, how mucho would be the cost for this system and API's on a montly basis? like a total for this example? Suscribed and grateful!
Hi Stephen, I am a tech youtuber in Korea. I found this video incredibly engaging and love to share this video to my Korean youtube subscribers. I was wondering if I could kindly request your permission to use a portion of this video in one of my upcoming videos. Of course, I would properly credit you and provide a link to the original video in my content. My intention is to briefly explain how you use MAKE while showing a segment of your video. Thank you.
Good morning, completely new to this, so forgive the remedial question, but with the code you provided one can copy and paste to receive same completed results? Change the artice to be specific to my needs? Do you offer freelance services? Thanks
If I want to change the final result to the size of the video that fits the shorts or reels, do I just need to modify the height and width of the output and image in the "Story Types" table? Then What should be the number of the four items: 'Output width', 'Output height', 'Image width', and 'Image height'?
Hi Stephen I'm just following up this tutorial will be great to have a flow chart to understand better the structure of the tables maybe is not the goal of the video but it will be nice to understand the structure to better see the connections for now I'm just following up maybe at the end I will make the connections
I haven't done a full calculation but my guess is ~100/month if you were making a ton of them. Not a lot for a single video but if you make a lot of them.
@@StephenGPope Glad he asked this question. Excellent video! Not to sound anal, but what is 'a lot' in terms of 100/month? Like a specific number range of videos that this will make for that amount
around a min of 100$ per Month (around 200 min of Videos per Month) (max 300$ if you create a lot of videos) for the Tools inclung gpt4 and freaking lot of time to copybuild this stuff, just finnished it, took me around 6 hours
This is epic!! only problem is I got 17:30 mins into this and realised when it came to adding the new section its part of the teams and it cant be done on the free mode, which is bummer but hey im gonna have to upgrade to see this through, jus wort ha mention is all :)
Hey guys, started building the module but got some problems when it is about chatgpt, where I get this error message: "You have consumed all credits in your OpenAI account, or exceeded your monthly OpenAI budget." does anyone know what plan I should get for chatgpt or another solution?
I understand how you use Leonardo to generate the images and GPT for the transcripts. but how ddi you turn the images into visdeos the text into audio and then combine everyything without a video editor?
Hi Stephen, are 100% AI content generation blueprints part of kontent engine db extreme? In other words is it possible for that product to generate AI only content?
Great tutorial! Unfortunately the "Run Script" functionality from AirTable is now behind the Teams Paywall. Is there any possibility for a workaround for this part?
AI automation HOW-TOs + Community → www.skool.com/no-code-architects
Easily distribute 100+ pieces content/week → kontentengine.io
Could you please create a video on how add custom text as well,on top of these videos?
OMG, this is awesome, you blow my mind. Thanks a lot for sharing it's a huge job, but easy to understand.
Spent about a day following this tutorial step by step. It took a while for me as I'm new to this but I got to the end and got a perfectly functioning automated video creation system up and running. A few thoughts for anyone who's serious about doing this: expect to spend an initial $100 or so for all the APIs. As of September 2024, I paid $50 for 120 minutes of json 2 video (i think it was their cheapest option), and ran out of make's free runs pretty quickly so had to buy a subscription for that as well. As soon as I switched to leonardo's animation mode, it ate up those points really fast - about $5 per video of about 200 words, it was generating 2-3 clips per sentence - still trying to figure out how to reduce the amount of clips generated for each story. Overall I'm really impressed with this new way of content creation. Anyone who wants to generate lots of content rapidly and doesn't mind the inconsistency of AI generated art can certainly benefit from doing this. Stephen, thanks for putting together such a great tutorial and sharing all this knowledge for free.
hey bro what do you mean by this = how to reduce the amount of clips generated for each story ???
@HaydenLee , can we have any updates? Where I can find you?
Has the revenue generated been more than the subscriptions?
@@socloseagency I think he wants to find the right prompting so he can have a good result with his images with one try. This is a nice video, but it could have been better with a disclaimer that it costs a lot of money (> 100 dollar) per month to make multiple videos with this methods. You need a subscription with Airtable to run the scripts, with openAI for the API, with Leonardo to AI the images, with Eleven Labs for the audio etc etc.
@HaydenLee you can edit the Story Types prompt, where it says to generate an array of 3 strings. If you make that 1 string, you will get 1 image per scene.
Introduction and Overview (00:00:00)
Setting Up the Airtable Base (00:00:56)
Building the Story and Scene Tables (00:04:26)
Generating Story Types and Prompts (00:06:08)
Linking Stories to Scenes (00:09:04)
Creating the Video API Table (00:11:29)
Generating AI Image Prompts (00:16:24)
Generating AI Images and Videos (00:21:54)
Building Scene Audio and Video (00:47:22)
Merging Audio and Video (00:58:12)
Building the Final Video (01:12:22)
Testing and Automation Tweaks (01:24:09)
Using Animated Videos Instead of Stills (01:31:12)
Final Testing and Conclusion (01:37:10)
This channel will blow up. Wild stuff. thanks for sharing.
wow thanks for the support!!
You’ve been elevated and promoted to ‘’GENERAL’’ you’ve outgrown your previous position 🎉 CONGRATS AND THX , CARRY ON AND CONTINUED SUCCESS , AN ABUNDANCE OF SUPPORT IS DEFINITELY AFOOT.
hahaha, thank you!
Here's a detailed breakdown from the transcript of the steps to build the automated faceless video creation system:
Set up the Airtable base:
a. Create tables:
Stories: Main table to manage overall video projects
Scenes: Breakdown of each story into individual scenes
Story Types: Define different types of stories and associated prompts
Video API: Track media generation requests and results
b. Set up fields in Stories table:
Story ID (Auto Number)
Story Type (Link to Story Types)
Scenes (Link to Scenes)
Source (Long text)
Generated Story (Long text)
Ready to Go (Checkbox)
Done (Checkbox)
Video (Attachment)
Number of Scenes (Formula)
Completed Scenes (Formula)
Ready (Formula)
c. Set up fields in Scenes table:
Scene ID (Auto Number)
Story (Link to Stories)
Scene (Long text)
AI Image Prompts (Long text)
Videos (Lookup from Video API)
Audio (Attachment)
Combined Video (Attachment)
Merged Audio and Video (Attachment)
Number of Videos (Count)
Completed Videos (Count)
Ready (Formula)
d. Set up fields in Story Types table:
Story Type
Story Prompt Text (Long text)
Scenes JSON Prompt (Long text)
Output Width (Number)
Output Height (Number)
Image Width (Number)
Image Height (Number)
e. Set up fields in Video API table:
Job ID (Text)
Video (Attachment)
Scene (Link to Scenes)
f. Create views for different stages of the process:
Generate Story
Generate Scene Videos
Build Scene Audio and Video
Merge Audio and Video
Build Full Video
Configure Make.com automations:
a. Set up webhooks:
Leonardo AI image generation callback
Leonardo AI video generation callback
Json-to-video scene completion callback
Json-to-video full video completion callback
b. Create main automation flows:
Generate Story and Scenes
Generate Scene Videos
Build Scene Audio and Video
Merge Audio and Video
Build Full Video
Implement the story generation process:
a. Trigger from Airtable when "Ready to Go" is checked
b. Retrieve story details from Airtable
c. Use GPT API to generate full story based on source and story type prompt
d. Update Airtable with generated story
e. Use GPT API to generate scene breakdowns and AI image prompts
f. Create records in Scenes table for each generated scene
Set up image/video generation:
a. Trigger for each scene without associated media
b. Use Leonardo AI API to generate still images or animated videos:
For still images: Make one API call
For animated videos: Make two API calls (still image, then animation)
c. Update Video API table with job IDs and media URLs
d. Link generated media to corresponding scenes
Implement audio generation:
a. Use 11Labs API to generate voiceovers for each scene
b. Save generated audio to Google Drive
c. Create shareable link for the audio file
d. Update Scenes table with audio file URL
Create scene videos:
a. Use Json-to-video API to combine images/videos with audio for each scene
b. Configure video parameters (FPS, dimensions, zoom effects)
c. Set up webhook for completion notification
d. Update Scenes table with combined scene video URL
Assemble final video:
a. Trigger when all scenes for a story are ready
b. Retrieve all scene videos for the story
c. Use Json-to-video API to combine all scene videos
d. Add captions to the final video
e. Set up webhook for completion notification
f. Update Stories table with final video URL
Set up error handling and retries:
a. Implement "Break" modules in Make.com for rate limit handling
b. Add error filters for specific error codes (e.g., 429 for rate limits)
c. Configure retry attempts and intervals
Create and configure Airtable views:
a. Set up filters to manage workflow stages
b. Create views to trigger different automation stages
c. Design views for easy monitoring and management of the process
Test the entire process:
a. Run test cases with different story types and sources
b. Monitor each stage of the automation
c. Verify output quality and consistency
d. Identify and fix any issues or bottlenecks
Optimize and refine:
a. Adjust prompts and parameters for better output quality
b. Fine-tune automation timings and error handling
c. Optimize Airtable base structure and formulas if needed
Optional: Modify for animated videos
a. Update Leonardo AI API calls to generate animated videos
b. Modify Json-to-video configuration to handle animated inputs
c. Adjust timing and transitions in final video assembly
Documentation and maintenance:
a. Document the system architecture and workflow
b. Create user guides for operating the system
c. Set up monitoring for API usage and quotas
d. Plan for regular updates and maintenance of the system
This detailed breakdown provides a comprehensive guide to building and implementing the automated faceless video creation system, allowing for customization and expansion as needed.
niiiicee! thanks!
@TravisChristopher did you get the automation to work 100%?
Btw I did everything on video but there are also additional costs.
- Airtable 24usd
- Leonardo ai 9usd
- eleven labs 5usd
- json2video 50usd (if you dont want have a watermarks on videos)
yes, thanks for adding these!
you can do it anywhere completely on free versions.
@@WojciechHerman What do you mean?
@@WojciechHerman how
you forgot about chat gpt
What I learned from this video:
- Automating the creation of faceless videos with animated scenes, voiceovers, and captions.
- Using separate tables in Airtable to manage stories, scenes, and image prompts.
- Defining story types to generate different outputs, like comic news or photorealistic images.
- Triggering automation to generate and assemble stories and scenes.
- Creating image prompts and animated videos for each scene part.
- Generating audio and assembling it with video parts into a final product.
- Configuring Leonardo API for image generation with webhook callbacks for updates.
- Using webhooks to link Json to Video’s response to Airtable.
Thank you Stephen this is some great content, keep going!
Right great summary and thanks for the feedback.
@fig100 did you get the automation to work 100%?
Stephen, I follow you from Bolivia (South America) and I love your video, I have been waiting for it for a long time, for me this is pure gold and thank you for giving us so much FREE quality content. Que Dios te bendiga
I'm so glad it was helpful! thanks so much for your feedback!
Here from Venezuela, this is gold!
Lol, everyone knows where Bolivia is 😅. Don't need to tell it's in South America.
This is absolutely the first time that I blindly subscribed to a bunch of tools just to see a project through, you are the man Stephen. You should definitely do this more often, this channel will blow up for sure.
00:01 Automate building faceless videos with Make and Airtable
01:54 Automated process to create videos from stories
05:28 Setting up database fields for storytelling automation
07:17 Setting up image and video dimensions
11:12 Creating a table to store images and videos generated using the Leonardo API
12:56 Creating and managing video-related fields in the database
17:00 Automations and scene assembly process
18:52 Adding new columns and creating formulas in a storage table
22:21 Creating an automated process for generating AI-generated videos.
24:01 Ensure specific rows enter the view
27:19 Automated process to generate stories using record data and ChatGPT
29:04 Updating permissions and executing tasks with FACELESS
32:27 Creating a new row in the scenes table linked to the story for each scene generated from ChatGPT.
34:03 Automated process for generating AI images and videos
37:36 Automating video generation process with script and automation testing.
39:05 Setting up Leonardo module for generating images
42:30 Automating video creation and data entry in Airtable
44:09 Updating records with video and image URLs
47:53 Setting up web hooks and automating scene audio and video generation
49:31 Creating and using an iterator to loop through images and create JSON
52:59 Setting up video structure with necessary IDs and parameters.
54:48 Creating a web hook in Airtable for Json endpoints
58:12 Automating speech synthesis with 11 Labs
1:00:00 Automated process of updating audio and merging audio with video
1:03:40 Guide on creating scenes using JSON for video and audio elements
1:05:34 Using web hooks to automate processes with Airtable and video rendering
1:08:52 Troubleshooting and optimizing automated scene generation process
1:10:23 Automated process creating video objects and scenes with delays for successful completion
1:13:48 Automate video creation process with FACELESS Videos
1:15:34 Processing and parsing Json elements for output
1:18:59 Customizing video elements using specified settings
1:20:33 Updating the main story and testing the trigger
1:24:08 Transitioning from still images to animated videos using Leonardo
1:25:53 Setting up automation for processing video content
1:29:05 Troubleshooting the automation process
1:30:40 Automated video generation and data integration
1:34:05 Transitioning from static images to automated videos
1:35:40 Automated generation of scenes, images, and videos
1:39:09 Automated system for story types and image prompts
Thanks
Thanks, Stephen! I followed all the steps and successfully completed the entire automation process. You're truly amazing!
Hello Ahmad. Could you please share with us how much it costed you to stand this automation?
This is such an impressive workflow! The way you designed each table to break down scenes and different story types is brilliant. I especially loved the quote, 'It just gives me a lot of flexibility to change the output of the video and the style without having to go into the automations and tweak all of these individual nodes.' Have you considered including an automated caption styling option to add even more customization possibilities?
And if you're looking to enhance your visuals quickly without the hassle, Aux Machina might be the perfect companion tool rather than Leonardo. It’s a super easy, intuitive way to generate realistic photos, so you can focus on creative freedom without getting bogged down in complex settings.
Damn what did I just watch ??????
This is mind blowing thanks a bunch Stephen. I enjoyed every little bit.
First time in my dev journey I subscribed for tools just to get a tutorial done. Thanks
💪
I just watched this video in its entirety, and it is definitely the best video of the year. Thank you so much, Stephen G. Pope, for everything! Greetings from Brazil.
🙏🙏🙏
🐀
Wow, i am going to be a member of your group monday. 5 week vacation going to be spent on learning this + more.
This just opened my mind and ideas are flowing in.
Excited, see you there!
Right on cant wait to help you out! Were going to be doing awesome stuff in the group!
Your method of delivering information is commendable, and I appreciate the quality of your work.
wow thank you very much, that means a lot to me.
Brilliant video. Something that would be really useful for marketers would be the ability to create these videos quickly but somehow pull footage and images from a repository of “approved” on-brand images / videos rather than using generative assets from Leonardo, stable diffusion, etc. doesn’t seem like it would be that hard to do but I’m guessing there would need to be a few alternative steps in make to search a DAM or storage location for assets with tags that match the script.
Yeah it wouldn't be too hard, especially using AirTable.. keep it simple like two keyword query, and you can have it pull random relevant content
yea you could definitely create a pool of content to pull from, could even tag it so the system would know what to pull from, great idea!
hi @EdwardAustin, i read that you have finish the tutoriel so i would like to ask you if you can help me because i cant do more after 1:06:42. When i start the automation with my last module and other side i click on test the action (airtable), they says : The scenario run was completed. but nothing its happen ( i dont have notification on my HTTP module). its hard to explain because im beginner but maybe you could help. Thank you
Hi,
I'm stuck on the video and audio merger. I followed your steps and used your prompt, which generated 9 scenes. For the first scene, I got 3 images and created a 15-second video. However, when merging audio and video, I used -2 for video and -1 for audio, resulting in only two images being shown. I saw a comment suggesting to use -1 for the video, which gave me a full 15-second video, but the audio cuts off at 7 or 7.5 seconds. What am I doing wrong? Can you help?
Amazing. The AU exchange rate is going to be a barrier for my subscription right now. Love your work.
thanks for your support and feedback! I can help you in my free community as well skool.com/content-academy
Wow! Guess I am going to wait for this to sink in to my slow brain 😅
I am very grateful Stephen for making this complex work simple and thanks for the wonderful explanation.
👍👍
Thought it was stopping at the intro then he be like, now I show you All. Holy cow. You're a God
💪
the same here
Thank you! With this video, I am truly amazed. I expect that anyone with automation skills has found an online treasure trove. You can sell your skills or use them to earn money. Thanks again-this is a game changer!
thanks for the kind feedback! it means a lot!
I've watched this 3 times. It's a big confidence booster.👍💪
Just need to come up with a syllabus on learning. Checking out your skool community to start the journey. 👍💪🤯
right on!! thanks for that feedback!
@@StephenGPope Every time I re-watch this it makes more sense.
Having a bit of trouble connecting my domain to Google business. Doesn't want to verify the dns text file.
I think it's because blue host didn't have a 4 hr option. Once that's done I can get "Make" and airtable.
Wow 😮 I'm realizing: I don't know what I don't know... 😕
@@livingworkingoutsidebox did you make it work 100% mate?
This is sooo good... thanks heaps Stephen. It took a bit to sort out the google drive connection, but a short video on youtube sorted that out.
Love all your content... so much to learn! 😊
He said to thank him you can vote labor into office.
Glad you got it!
I love this. I just joined the community. I wanted to share that if you have timestamps on your youtube videos that would be a huge help to going back and forth to specific tasks/parts
We can actually do that ourselves as commentors, just like now a timestamp and label it such as:
00:00 - Start
15:00 - Mid
1:00:00 - End.
Thank you for that feedback and welcome to the community!
waww hands up bro ( honestly people have to pay for this )
I just finished the video and managed to create my first automation video (I set the speed to 0.5)
the cost will be about 150 dollars for 200 minutes (very very cheap)
i will now try to adjust the tiktok video and find a way to add a lmusci background
thanks again Stephen with all my respect for you and your work
🙏🙏
what do you mean the cost was $150 dollars for 200 mins???? thanks
@@StephenGPope I’m having trouble figuring out how to automatically upload the final video to Google Drive after it’s generated . Could you please explain how to do this?
@@saulbelford7618 the cost of APIs be careful when you use Leonardo API videos because it's cost a lot )
@@perfecto7834 did u find out?
Hey Stephen, great tutorial, especially for a beginner like me. Its my first time i use make and I implemented it successfully. Thank you.
how long did it take u?
When you start on the second part of the automation in make , when i created the airtable get record and selected scenes and want to add a filter the action keeps pointing out to generateStory instead of generateScenesVideos - Why is that so??
Hi Stephen, No words and mind-blowing tutorial. Thanks a lot for taking the time and explaining step-by-step. The most useful and highest quality video
Hey thanks glad you liked it 👍
This was the best BOOTCAMP for Airtable + Make + others AI tools......Unfortunately I cannot give a Like twice ;)
🙏🙏 thank you for that feedback it means a lot
Thank you for such valuable information and taking the time to teach it. Is it possible to use Google Sheet instead of AirTable?
I can make it in midjourney?
Yea i did that video last week check it out
Simply mind blowing. Pictory, invideo going to close shop soon hahha. BTW, about how much to invest in for all these tools mentioned? Thanks
= ) the cost depends a lot on your specific use case so I don't have an estimate
@StephenGPope are you interested in selling this automation as put together currently to a user like myself?
thank you for sharing, what app you use for make arrow pointer on mac?
Same question I have…
demo pro on the mac
demo pro on the mac
I learned a ton from this and it really opened up my mind. Thank you Stephen.
❤ Your knowledge. The pace and details. Lovely. Awesome video ❤
🤝 thank you for that feedback!
Re: the JSON2VIDEO errors, I was getting "Error: Movie duration cannot be zero" on the generateSceneAudioVideo action until I realized I'd missed checking the "elements" box on the Array Aggregator at 52:31. JSON2VIDEO thought my video was zero seconds because it wasn't receiving the necessary info. Just one more thing to check when troubleshooting this issue.
This was an awesome video, thank you for the content. Question, is this setup inside of the program, or would have to watch the video and go step by step again in order to implement it?
Its still better to watch the video but you get access to tech support and calls. Next month ill release a new one that is more plug and play.
Great video! What tool do you use for the purple arrow demonstration? It's so cool!
Thank you so much. That was awesome. I can’t believe how fast you talked. It’s not a bad thing I can slow you down. And watch it again, but I can’t imagine go yourself and drink of water 😂 like an hour you were like an auctioneer.😂you might be really really good at editing and I didn’t catch it, but you didn’t mess up at all like a perfect marathon sprint checked out your school that’s cool. I’ll hang out there but can’t afford it yet.
Haha oh man this was heavily edited and i had to take a ton of breaks. Thanks for the comments! It took me ALL day. It was really hard lol
This is absolutely fantastic! Thanks Stephen.
💪💪
This video deserves 10 million likes
🙏🙏🙏🙏🙏
That’s seriously badass presentation ! I think i’m going to have some fun with python this weekend
Can I use this for long form content? My content always comes out as 55 seconds long?
You absolute hero. That was brilliant.
hey thank you! that means a lot to me! 💪
Are you MAD bro! This is insane 🤯🤯. I can't wait to find a day to watch this video at 0.25 speed and follow all the steps to recreate this automation.
Im a madman! 😀
Should I adjust the speed from .75 to .25???😂
@vinceslas did you get the automation to work 100%?
@@shawnjohnson4253 No I ran into multiple hurdles and couldn't complete the whole process. I just need to put my head around it.
Thanks for the great video! I appreciate the detailed . Do you recommend any other API services for generating different motion effects?
On sunday i have a new video coming out on lumalabs which is pretty dope
What kind of automation do you teach in your community? What applications of AI? can you give examples? I can see you're producing a short film here, so I'd like to know how this info is applicable to other businesses. Thanks!
yea, have a getting started Make.com + Airtable automation as well. Plus calls.
Hi Stephen, thanks for this! I am wondering why not use Dalle 3 or Midjourney (routed through make) instead of Leonardo for the image generation part? Is it because of some price or quality reason? Then you could always send it to Leonardo later for the "movement" part. Appreciate your thoughts!
Thats a good idea. You could do that sometimes it comes down to preferences
Wow just wow! can you share the make scenario?
Skool.com/no-code-architects
i keep getting this error - in the merge audio + video - this error comes from JSON2Video Error: Scene #1, element #1: The URL provided is not a video file and cannot be used as a source for a video element. The URL provided is an image file (codec: mjpeg). anyone knows why so ?
Cool, thanks for sharing, when you say reply about cost 100$/month for making many videos? what's the average for "many" videos? just to get some numbers. cheers!
I really don't have exact numbers for you sorry. you'll need to estimate it by going to the sites and looking at what it will take, and how long the videos are, etc. Not a simple problem to calculate.
I was wondering the same thing. I think if you are going to go large scale, this would proabably be well worth the $$. I'm going to try and figure out a $/min cost.
@@TinkerandBuild Hey! Have you figured out the $/min cost already? would be very nice to know. Thanks in advance.
@TinkerandBuild have u figured it out?
Any updates on the minimum cost?😁@@TinkerandBuild
God video, I have a question. It seems like you already know what to do because you've made it work and are now creating the video. But I'm curious, what does your process look like when you're actually figuring things out?
Some people have asked i will try and make a video on that.
01:12:43 no Sir YOU are awesome I’m following along easily thanks to your stunning attention to detail for absolutely every single step.!
Wow, thank you!
@@StephenGPope I finished my first one yesterday. Having trouble with the prompt for the story not updating to the new source but I’m confident the answer is at the beginning lol. Thanks so much for your effort.!!
awesome video with explanation of each item. thanks.
happy to! thanks for your feedback!
Curious how many operations it takes to make a video, how much the video costs in terms of eleven labs, LLM tokens and the monetization strategy for the video. I saw a SAAS company working on something like this but generating marketing videos for B2B clients
Who was the saas company? Looking for solutions that do this and considering the pros and cons of rolling my own vs buying.
I haven't calculated the number of operations exactly, probably 100+ operations per video
@stephengpope How much would you approximate it would cost with your automations for a 2 min video? Also, do I need to know how to code to put this entire flow together?
Can we replace Airtable with some other open-source system running on our homelab infra?
I wish to try replace paid services with something which can run on my servers at home.
I'm sure its possible, maybe with nocodb.com
If the scenes generated describe a main character. How would you go about ensuring the characters visuals are kept consistent across all scenes. I have randomly generated the scenes from the story and got a slightly different result, which then describe a main character, and when those prompts are passed to leonardo, the main-character doesn't match visually from scene to scene. I think this issue is present in many story based generation. How would you make leonardo aware to keep the entire context consistent.
try repeating the character description on every prompt with detail.
try sending the first photo of the character do Dalle and asking it to describe the character, that way you can use its description as the promp to the others.
much appreciated for your informative video.
I have two questions:
Can we change it the language to something else?
Can we make the video doesn't exceed one minute?
This is fantastic! Thanks Stephen!
Dont thank me yet 😆 its a doozy!
With all these automations you could also create chapters for this video.
I actually have a video on that but havent started using it yet 😅
would this work for 10min+ videos too?
sure I don't see why not!
Nice Video!
what recording screen software do you use? :)
Ecam live and thanks!
need to hide this video from blowing up, cant let this nugget to be shared widely
it's FREE!
two questions:
1. what if I dont like the narration? does it allow me to edit it before it goes through the wholr process?
2. can I generate more images per secene, or shorten the 15 seconds for each secne? the tempo right now is too slow.
thanks.
1. It can yes just have to add a check to the filter that lets you approve.
2. Sure you can
Would you mind breaking down how much rev you're generating from content vs. info products? We don't need actual numbers just percentages I think would be the most useful.
Very little rev comes from content views and affiliates maybe like 750 to 1.5k a month
Hello @stephengpope it is working perfectly!
Can you explain in detail how to add music on the last step ?
I tried different ways but i cant make it work in the same step, only by adding new step but loosing json2video quota.
Love you 👏😎😜
So how this works? for every type of video you want to create, you have to do the whole propcess from 0? and how long videos you create like this could be?
once you have the system in place you can generate videos in a few minutes.
what is the right way to make a youtube shorts with that system please?
what are the output width and height?
thanks!
1080x1920
Your work is very good, I just have one question, my video is longer than 60 seconds, which generates a json plan error, how can I limit my videos so that the duration is a maximum of 60 seconds?
You would need to ensure that by limiting the story length and scenes by tweaking the prompts in the story type table
@@StephenGPope Could you remind me which part of the video or step by step to make this adjustment?
Thanks for the detailed guide.
I got stuck at the BUILD FULL VIDEO stage at minute 1:16:00
I run the script successfully but in MAKE it doesn't show that it produces the videos in JSON and therefore in the array aggregator there is no option to select ELEMENTS.
Do you have any idea that can help me with the problem?
probably better to join my free or paid community to post the question with a screenshot so I can see what you're talking about. Debugging this via youTube comments is HARD.
Anyone know why at 35:44, at “action” it says “generateSceneVideos” but for me it still says “generateStory” this is also the same when making a filter on that module?
I have the same issue! can't figure out how he did that
that means you didn't trigger the other action from Airtable, or perhaps the same value is being sent from the same automation.
you just need to make sure you're triggering from the right Airtable automation with the right action.
@@StephenGPope but the generateSceneVideos action shows up on your airtable module on make.com before you even create the automation on airtable? am i misunderstanding?
simply mindblown right now.
niiiice!! 💪
Absolutely amazing! Thanks a lot for all this value! im glad that i found you, i was literally looking for some real life example like this!, may i ask some, how mucho would be the cost for this system and API's on a montly basis? like a total for this example?
Suscribed and grateful!
You probably end up spending ~ $100/month to produce quite a few videos
@@StephenGPope thanks!
Is the integration with the Airtable AI worthwhile? I imagine if their generations are good enough, it would simplify the ChatGPT integration.
Good idea but havent tried it yet.
What a journey !
Could I ask how much you pay per month for all this tools ? (airtable/make/Leonardo/chatgpt/jsontovideo/ etc )
Aproxx 100/month depends a lot of what you do
Hi Stephen, I am a tech youtuber in Korea. I found this video incredibly engaging and love to share this video to my Korean youtube subscribers.
I was wondering if I could kindly request your permission to use a portion of this video in one of my upcoming videos.
Of course, I would properly credit you and provide a link to the original video in my content. My intention is to briefly explain how you use MAKE while showing a segment of your video.
Thank you.
Most inpressive make I've ever. seen insane thanks, wjat wad tour background to learn to do this or how long to learn, people would pay for this
thank you! I'm a content creator and also a software engineer for 30 years!
I must say this is like out of this world
🤘🚀 honestly tha ks for your feedback
Good morning, completely new to this, so forgive the remedial question, but with the code you provided one can copy and paste to receive same completed results? Change the artice to be specific to my needs? Do you offer freelance services? Thanks
yes, once you have the system you can tweak it to work with your articles, image type, etc! sometimes I do offer freelance services.
@@StephenGPope Great, whats the best way to contact you? Thank you!
If I want to change the final result to the size of the video that fits the shorts or reels, do I just need to modify the height and width of the output and image in the "Story Types" table?
Then What should be the number of the four items: 'Output width', 'Output height', 'Image width', and 'Image height'?
You are the BOSS, this is. gold nuggets! Thanks for sharing all this with us
hey thanks! appreciate that! happy to share.
Wow, thanks for this 🙏
Hi Stephen I'm just following up this tutorial will be great to have a flow chart to understand better the structure of the tables maybe is not the goal of the video but it will be nice to understand the structure to better see the connections for now I'm just following up maybe at the end I will make the connections
Almost positive that is in his paid group.
Wauu! Impressive 🔥 any idea of the cost to run this?
I haven't done a full calculation but my guess is ~100/month if you were making a ton of them. Not a lot for a single video but if you make a lot of them.
@@StephenGPope Glad he asked this question. Excellent video! Not to sound anal, but what is 'a lot' in terms of 100/month? Like a specific number range of videos that this will make for that amount
How much total cost fir this whole automation? Including paid tools
A lot, which is why I'm guessing there's no response here
Waiting for the same answer
All of it.
around a min of 100$ per Month (around 200 min of Videos per Month) (max 300$ if you create a lot of videos) for the Tools inclung gpt4 and freaking lot of time to copybuild this stuff, just finnished it, took me around 6 hours
That is not the issue though. The result are mediocre at best 😂
This is epic!! only problem is I got 17:30 mins into this and realised when it came to adding the new section its part of the teams and it cant be done on the free mode, which is bummer but hey im gonna have to upgrade to see this through, jus wort ha mention is all :)
let me know how it goes = ) CHALLENGE ACCEPTED! get it!
Bro. How did you even figure this out!?
It came in a dream 😆
@@StephenGPope 😂😂😂
What software do you use to record the videos & the "pointer" to annotate?
hey there! it's called Demo Pro on the mac.
This is awesome video. Please can you tell me what's the average cost of making this automation for 30 days? 🤔
Depends a lot of how long the videos are and how many you make estimate 100 /month
Hey guys, started building the module but got some problems when it is about chatgpt, where I get this error message: "You have consumed all credits in your OpenAI account, or exceeded your monthly OpenAI budget." does anyone know what plan I should get for chatgpt or another solution?
I understand how you use Leonardo to generate the images and GPT for the transcripts. but how ddi you turn the images into visdeos the text into audio and then combine everyything without a video editor?
Leonardo turns images into videos too
Hi Stephen, are 100% AI content generation blueprints part of kontent engine db extreme? In other words is it possible for that product to generate AI only content?
hi there, can you share which toll are you using to draw the arrows?
Demo pro on the mac
@@StephenGPope thank you ! great content btw i'm really enjoying your channel :)))
Great tutorial!
Unfortunately the "Run Script" functionality from AirTable is now behind the Teams Paywall. Is there any possibility for a workaround for this part?
NOT easily
Is it possible to do the same but to create long form videos like 30/40/50 minutes long?
Sure!
Can this be done with fre airtable plan? Or google sheets?
it could but it would waste a lot of make operations using their on demand triggers, you'd spend more on Make than the savings for Airtable.
Is there a way to then take the completed video and then auto post them to RUclips Shorts/ Tiktok on a given schedule?
Sure absolutely lots of ways
Do you have the option to also add background music to the videos?
Sure