Data-driven video is video where a data source, such as a spreadsheet, a JSON feed or an API, decides what each version of a template says and shows. One designed template plus one row of data makes one finished video, and a thousand rows make a thousand videos. For After Effects teams, Templater is the plugin that does the pairing, QUE can hold the data and the job queue, and Projector can render and deliver the results.
That is the short version. The rest of this page is the long one, written for the producer, developer or marketer who has a pile of versions to make and wants to know how the pieces fit. It covers what the workflow looks like, which software does which job, how to go from a data source to finished files, and where the approach is the wrong tool.
What data-driven video is and how it works
A normal video is one finished thing. Change the name on it, swap the product, translate the headline, and a person has to open the project, make the edit and export again. Do that three times and it is fine. Do it for every customer, every product in a catalog or every language you ship in, and the editing becomes the whole job.
A data-driven video flips the order of work. A designer builds the template once. The variable parts, the text, the images, the footage, the colors, are marked as slots. Then a table of data fills the slots, row by row. The designer’s time goes into the design. The repetition goes to the software.
Four parts make up every pipeline of this kind, whatever tools you pick.
The template. In our world this is an After Effects project with a master composition. Its layers are rigged so they can accept outside content. A template that was never built to take data will fight you the whole way, so the quality of the rigging matters more than the quality of the plugin.
The data source. Templater works with Google Sheets, tab-separated files exported from Excel, Numbers or Sheets, and JSON files or feeds. QUE is a cloud-based, JSON-based data source built to work with Templater, and it has an API for apps that want to send data in. Our page on data sources in Templater covers the differences.
The merge. This is where each row lands in the template. Text goes into text layers, images and footage get swapped, and layout rules adjust each dynamic layer against its container and its sibling layers, so a longer name does not break the design. Time Sculpting rules can crop, trim, shift and stretch layers, so footage of different lengths fits the design instead of bending it.
The render and delivery. Each merged composition becomes a file, or a project copy waiting for further work. Someone or something has to start that run, track it, and put the finished files where they belong.
The Pipeline
Four parts every data-driven video needs
-
The templateAn After Effects project with a master composition whose layers are rigged so they can accept outside content.
-
The data sourceTemplater works with Google Sheets, tab-separated files exported from Excel, Numbers or Sheets, and JSON files or feeds.
-
The mergeEach row lands in the template: text goes into text layers, images and footage get swapped, and layout rules adjust each dynamic layer against its container and its sibling layers.
-
The render and deliveryEach merged composition becomes a file, or a project copy waiting for further work.
Once you see the four parts, the confusing product names sort themselves out. Each Dataclay product takes one or more of these jobs.
A hypothetical example
Picture a training team at a company with new hires starting every week. The welcome video has the hire’s name, the office, the manager’s name and a short clip from the local site lead. Today an editor makes each one by hand. This is a made-up case. With a data-driven setup, the HR system exports a sheet with those four columns. The template has four slots. The designer rigs it once, and each new row produces one video. The editor now reviews a sample and fixes the template when something breaks, rather than exporting every file.
Choosing software for After Effects templates at scale
If you are asking which software generates data-driven videos from After Effects templates at scale, the honest answer depends on where you sit in the pipeline. Here is how our own products divide the work, with the trade-offs stated plainly.
Templater: the plugin that merges data into the template
Templater is a plugin that runs inside Adobe After Effects. It comes in three editions, and each one adds to the one before it.
Rig prepares a project for dynamic content. You map your data to composition layers and preview how the layers change as the data is processed. Rig is the right pick when swapping content by hand eats your time, when you need to prepare a project for another machine running Pro or Bot, or when you want to build a template to sell or hand to someone else.
Pro includes everything in Rig and adds batch rendering and batch replication of re-versioned master compositions from your data source. Pro fits a studio that has a client database and a quick deadline, or several artists working on variations of one master. A person starts each batch.
Bot includes Rig and Pro and takes the person out of the loop. It checks for incoming data and starts new renders by itself, or you can push data in from the command line. Bot is the edition for videos on demand, for apps that feed data in, and for very large data sets. If you want the full picture of the family, read the complete guide to Templater.
Templater Editions
Which Templater edition fits your job
| Edition | What it adds | Who it fits |
|---|---|---|
| Rig | Maps data to layers; previews how layers change | Preparing a project for another machine, or building a template to sell |
| Pro | Batch rendering and batch replication from your data source | A studio with a client database and a quick deadline |
| Bot | Checks for incoming data; starts new renders by itself | Videos on demand; apps that feed data in; very large data sets |
Pro and Bot offer a choice at run time. Render produces finished files. Replicate saves a copy of the project for each version, so you can hand-finish a version before it renders, or send the copies to the After Effects render queue or to Adobe Media Encoder. Which one you want depends on whether anyone needs to touch the versions between merge and render.
QUE: the data source with an API and a job queue
Spreadsheets work well until the data gets large, complicated or fed by another system. QUE uses JSON instead of rows, and it has a built-in API, so a web or mobile app can send in what a customer chooses and get a video back. It also has a Zapier integration for connecting to other apps without code.
QUE organizes the work into four parts. An Autograf is a template with layers rigged to accept data. A Job is the JSON object holding the data for one video. A Campaign is a collection of jobs. A Satellite is a computer running After Effects and Templater that processes the jobs. If you are weighing the two, our comparison of QUE and Templater lays out which handles what. For a low-cost way in, there is QUE Meter, and the QUE Meter announcement explains it. Our QUE API walkthrough shows the first job end to end.
Projector: parallel rendering and delivery
Projector is a desktop application for Windows and Mac. It renders several customized videos at the same time, using the After Effects rendering engine in parallel based on the number of cores in the machine. A machine with eight cores renders eight videos at once. It can transcode into more formats than After Effects offers natively, and it can send finished videos to Vimeo, JW Player and Frame.io, with AWS S3 for storage. Rendering stays on your own machine, so your assets are not visible to another party, including us.
Projector has to run on the same computer where Templater is installed and where its output files are stored. Read what Projector is before planning around it, and QUE versus Projector if you are unsure which one you need.
Autografs: templates that are already rigged
Not everyone who needs data-driven video has a motion designer on staff. Autografs are video composition templates in categories like real estate, automotive and sports, already rigged with Templater customizations. You add data, tweak the template for your audience, and use it. They are a fast start, and a custom template will fit your brand better. Our guide to Autografs has the detail.
Where our tools are not the answer
Templater needs After Effects. If your team does not use After Effects and has no plan to, a tool built around a browser editor will be a more natural fit, and we would rather say so than pretend otherwise. We also do not generate a video from a text prompt. Templater fills templates a person designed, which is why the output matches the brand exactly. Our post on generative AI and data-driven templating explains where the two approaches differ.
From a data source to finished videos
Most questions about this subject come down to one problem. How do I get from a table to a folder of finished files without hand work in between? The steps below answer that in order.
Connect After Effects to a data source
Start with the data, not the design. Decide what changes from video to video and give each of those things a column. Keep one row per finished video. A name column, a headline column, an image path column and a language column make a fine first sheet. Templater reads the text as it appears in the sheet, so commas, quotes, apostrophes and new lines are accepted, and right-to-left scripts are supported. The installed fonts need to have the glyphs you require.
Next, rig the template. In the Templater panel you link text, image and footage layers to your columns. Then preview. Templater lets you preview how your compositions look with the data before you render or replicate, and that is where you find the name that is too long or the photo that is the wrong shape. The walkthrough on setting up and using Templater inside After Effects goes through each step.
Before You Run
What to do before your first batch render
- Decide what changes from video to video and give each of those things a column
- Keep one row per finished video
- Rig the template: link text, image and footage layers to your columns in the Templater panel
- Preview how your compositions look with the data before you render or replicate
- Check a sample, spot checks create still images or movie clips of specific frames from the data-merged compositions
- Test with your real data so you see how your own template behaves
Render many versions automatically
With Pro, you start a run and Templater works through the rows. With Bot, a data source with new rows is enough to start the work, with no one at the keyboard. In both cases the batch runs from the data, and you can name each output from a column in the same row, so the files are easy to find afterward. Our post on rendering versioned video in a batch shows the settings.
Before a big run, check a sample. Pro and Bot include spot checks, which create still images or movie clips of specific frames from the data-merged compositions. You mark the frames with named composition timeline markers, or Templater samples frames at even intervals across the composition. It is a faster way to catch a bad row than watching every output. Read preview modes and spot checks for the details.
Automate video creation with an API and dynamic data
If another system should trigger the video, there are two routes. The first is Templater Bot’s command line, which lets you push data in to create renders. The second is QUE, where your application sends a job through the API and a satellite machine renders it. The QUE API documentation is at que-api-docs.dataclay.com, and our post on sending render jobs with the API is the practical start. A job moves through a render status as it goes. It is ready, then queued, then processing, then done, and a job that fails shows an error code instead. The page on what each render status means explains the codes.
Personalized videos from data
Personalization is the case where each row is a person. The row holds a name, a plan, a city or a purchase, and the template turns it into the words and pictures that person sees. Three habits keep it working. Put only the fields that change into the data. Give every field a fallback in the design, so a missing value does not leave a hole. And review a sample of finished videos, not only the template. Our guide to making hundreds of personalized videos from data works through a full example, and personalized video from a CRM export covers the export side.
Localized videos from a translation sheet
Localization uses the same machinery with a different column. Each row is a language, and the text layers read from it. The design problem is that translated text changes length, so the template needs room and layout rules that cope. See using Templater for localization for how teams set it up.
Motion graphics from data feeds and CSV files
Content teams that make charts, scoreboards, price cards or news graphics from a feed use the same pipeline. The feed is the data source, and the template is the graphic. A tab-separated export from a spreadsheet works, and so does a JSON feed. The post on using a URL feed or JSON file as a data source shows how.
What about data collection and dynamic visualization?
People sometimes ask for a single video tool that collects data, processes it and draws it. Our products start after the collecting. Templater and QUE take data that already exists, from a sheet, a feed or an app, and merge it into a template a designer built. The visualization lives in the design. If the number should become a bar, the template has a bar layer, and the data sets how long it is. If you need a system to gather and process the data first, that is a separate piece, and our custom development team can talk through building it.
When this approach is wrong, and the bottom line
Data-driven video is not worth building for every project. If you need five videos and each one needs a different idea, hand editing is faster than building a template. If the design changes every week, you spend all your time rerigging. The approach pays when one design carries many versions, and when the differences between versions fit in a table. For a plain-spoken look at the tipping point, read when it is worth automating your After Effects templates.
The other trade-off is the setup. Rigging a template well, cleaning a data source and testing a sample all take real hours up front. The payoff comes on the second batch and the tenth, not the first. If you can try before you commit, do. Templater has a free trial that contains all features for all versions, with a watermark on dynamic layers, and QUE has a free version. Your pilot should use real data, because that is where the surprises live.
So here is the verdict. If your team builds in After Effects and has many versions of one design to ship, Templater is where to start, QUE is the next step when the data gets big or an app has to feed it, and Projector is the step when rendering and delivery become the slow part. If you are somewhere else, we would rather you pick the right tool than ours.
Here are answers to some questions teams ask when they are starting out.
Frequently Asked Questions
What is data-driven video?
Data-driven video is video built from a template and a data source. A designer builds the template once and marks which parts change. A spreadsheet, JSON file, feed or API supplies the values for each version, and software merges each row into the template to make one video. The result is many videos that share one design and differ in names, text, images, footage or language.
What is the best software for generating data-driven videos from After Effects templates at scale?
It depends on your pipeline. Templater is our plugin that merges data into After Effects templates, with Pro for batches and Bot for automatic runs. QUE adds a JSON data source, an API and a job queue. Projector renders in parallel and delivers files. Teams that do not work in After Effects may be better served by a different kind of tool.
How do I connect After Effects to a data source and render many versions?
Rig your template with Templater, link each dynamic layer to a column in your data source, and preview the result. Use Google Sheets, a tab-separated file or a JSON file or feed as the source. With Pro, start a batch and Templater renders or replicates each row. With Bot, new data starts the work automatically. Spot checks let you inspect sample frames first.
How do I automate video creation with an API and dynamic data?
Two routes exist. Templater Bot accepts data through its command line, so another system can push rows in and start renders. QUE has a built-in API, so your application can create a job with its data and a satellite machine running After Effects and Templater renders it. The job’s render status tells you whether it is queued, processing, done or failed.
How do I make personalized videos from data?
Build a template with slots for each personal detail, such as a name, city or product. Put one person per row in your data source. Templater merges each row into the template and renders one video per person. Preview and spot check a sample before the full run, and design a fallback for every field so a missing value does not break the layout.
Can I use CSV or spreadsheet data for motion graphics?
Yes. Templater reads Google Sheets and tab-separated files exported from Excel, Numbers or Sheets, so a CSV can be converted to that format and used. It also reads JSON files and feeds. Each row becomes one version of the template, so charts, price cards and scoreboards can come from a table without manual edits.
Can I try Templater before buying?
Yes. The free trial contains all features for all versions, with a watermark added to dynamic layers. The one other limitation concerns running a replication process inside the project file, where the option to send replicates to Adobe Media Encoder needs to be enabled. The command line is available in the trial. Test with your real data so you see how your own template behaves.
