Dataclay’s documentation does not name an exact moment when a live API feed becomes a liability in a render pipeline, so this page does not invent one. What the documentation does state is how a live URL feed works as a Templater data source, what The Bot can and cannot do with it, and which settings and error codes apply. Those facts show you where a feed needs extra care. The point at which a live feed starts causing failures in a particular pipeline is not given here, and the way to find it is to test your own feed with a small batch before you rely on it.
What a Live Feed Is in Templater
A Templater URL feed is a REST API endpoint that returns responses as JSON object arrays. You link it by entering its address in the Data source field in Batch Assets. In a Templater data source, each spreadsheet row or JSON array object represents a specific render job, so each object your endpoint returns becomes one job. Each column or property key represents a dynamic layer element in the video template, and property keys in a Templater URL feed must match dynamic layer names or reserved keys.
Other sources are available. A Templater data source can be a QUE campaign, a Google Sheets document, a TSV file, a JSON text file or a URL feed. All three Templater editions can use QUE, JSON file, Google Sheets, TSV and remote JSON URL feed data sources. A feed is therefore one choice among several, and the documentation treats it as a source of the same kind of data as a local file: JSON property keys are equivalent to spreadsheet column names in Templater.
The documentation is plain about the skill a feed needs. Setting up a URL feed as a Templater data source requires knowledge of configuring web servers. That is the first place a feed differs from a spreadsheet, because the endpoint is something you or your developers run and maintain. The guide on using a URL feed or JSON file as a Templater data source covers the setup.
The Render-Status Limit That Matters Most
The clearest documented limit on a live feed is about render-status. When The Bot runs, it follows a set routine. The Bot retrieves data source rows with a render-status value of ready. It updates all retrieved rows to queued. For each queued row, it sets render-status to processing, and then to done when output is finished. This routine depends on The Bot being able to write the new status back to the data source.
With a URL feed, The Bot cannot update the render-status key. The documentation tells users to register scripts to update it, and Dataclay advises custom scripts to update render-status values so ready data is not rendered again. In other words, if your endpoint keeps returning a row marked ready and nothing updates it, nothing in The Bot marks that row as finished. Rows that are still marked ready can be rendered again.
That is a documented reason to plan for the feed’s status handling before you turn on The Bot. A few related facts help you plan:
- Before enabling The Bot, the data source must have a render-status column with some rows set to ready.
- The value ready is entered by the user or by a front end application, and Templater writes the rest.
- With one Templater Bot, you or your app should only enter ready in render-status or empty the cell.
- The Bot will overwrite existing files if it renders to the same output file name, so repeated rows that share a name replace earlier output.
Before You Enable
What to have ready before turning on The Bot
- Data source must have a render-status column with some rows set to ready
- Enter ready in render-status or empty the cell, do not use other values
- Register a script to update render-status so ready data is not rendered again
- Property keys in the feed must match dynamic layer names or reserved keys
- Run a pre flight render of one video before relying on the full feed
Compare this with a spreadsheet. As of Templater 2.9.5, formulas in render-status cells are overwritten with their result when The Bot updates them, which shows that The Bot does write status values into sources that allow it. A URL feed is the source type where the documentation says it cannot. For details on settings, see the guide on configuring the Templater Bot.
Data Source Limits
URL feed vs. spreadsheet on render-status
- Can The Bot write render-status?
- URL feedCannot update the render-status key
- SpreadsheetWrites status values into sources that allow it
- Who keeps ready rows from repeating?
- URL feedRegister scripts to update render-status
- SpreadsheetThe Bot overwrites cells with result values
- Technical setup needed?
- URL feedRequires knowledge of configuring web servers
- SpreadsheetNo web server knowledge required
With a URL feed, The Bot cannot update the render-status key, so users should register scripts to update it.
How Often The Bot Reads a Feed
A live feed changes while The Bot is running, so the reading schedule matters. When you enable The Bot, Templater enters a server mode and checks your data source at regular intervals. Templater Bot’s default polling interval for new data is 0.5 minutes, or 30 seconds. For URL feeds and JSON sources, The Bot reads at intervals no shorter than 0.2 minutes, about 12 seconds. For Google Sheets the minimum is 0.3 minutes, about 18 seconds, because Templater uses the Google Sheets API and must regard API rate limits.
A live feed is read on a schedule, because The Bot checks your data source at regular intervals.
The Bot can also hold only so much at a time. Templater Bot can queue up to 20 rows or JSON objects at one time, which is both the default and the maximum. When you use more than one Bot on the same data source, the bot column value assigns a row to the Bot with the matching Bot Name, so the feed’s objects need that key as well.
Errors You Can Expect With Remote Data and Footage
When a feed or the files it points to cannot be reached, Templater’s error codes name the cases. Templater error code 6001 means HTTP request error. Code 1002 means error downloading a remote footage source, and code 1003 means Templater could not find the footage file in the network. Code 3001 means error conforming row data from a spreadsheet or TSV file. These codes are written to the log files, and the Templater CLI logs messages to templater.log and errors to templater.err.
Error Codes
What Templater error codes mean for remote data
| Code | Meaning | Where it appears |
|---|---|---|
| 6001 | HTTP request error | Log files |
| 1002 | Error downloading a remote footage source | Log files |
| 1003 | Could not find the footage file in the network | Log files |
| 3001 | Error conforming row data from a spreadsheet or TSV file | Log files |
Remote footage has its own events and options. Templater broadcasts Before and After Footage Download events around downloading remote footage, and Templater Footage Download fires for every remote footage download, including remote references cached locally. The Footage Processing and Footage Download event groups are only in Templater 3.4 or later. Templater Dynamic Essential Properties supports both local and remote footage paths for media properties.
If the footage a row points to is missing, Templater has settings for that case. The missing footage options are Ask me what to do, Skip the row, Use my branded placeholder and Cancel, plus Let me find the footage asset. Dataclay recommends the Use my branded placeholder option for smooth batch processing. The Footage group in Templater Preferences also has a checkbox called Consider empty cells or null values as faulty, and if empty cells are not treated as faulty, Templater holds over footage from the previously previewed row. The guide on passing asset URLs into a render job without exposing credentials covers the handling of footage links in more detail.
For Google Drive footage, the documentation notes several limits that also apply to any remote source of files. Streamed (Online only) Google Drive files that are not synced cause errors when Templater imports them, and some large Google Drive files might not download correctly through the Templater Media Connector. Dataclay advises enabling Available offline on all Google Drive files Templater must access. Personal Google accounts using the Templater Media Connector must reauthenticate every 7 days.
Designing a Pipeline Around a Live Feed
The documented facts point to a short set of design choices. None of them is a claim about when a feed fails. Each is a way to remove a limit the documentation describes.
First, decide who updates render-status. Because The Bot cannot write it back to a URL feed, register a script or have your application do it, so that ready data is not rendered again. Second, give every row a stable label. Dataclay recommends a column named id to label each spreadsheet row for Templater output, and the same idea applies to a JSON property key named id for a feed. Without ids, Templater uses the spreadsheet row’s index to generate unique identifiers. Third, match the keys. Property keys in a URL feed must match dynamic layer names or reserved keys, so check your endpoint’s field names against your layer names.
Fourth, think about the machines. Leaving Templater Bot enabled for long periods can slow versioning because of After Effects layer caching, and Dataclay recommends disabling After Effects disk caching on any machine where Templater Bot is left enabled. Fifth, use QUE when you want a managed queue. QUE is Dataclay’s render queue and API for sending jobs to satellite machines running Templater, and a batch of QUE jobs can be inserted with the Import Jobs API endpoint or from a file in the QUE Dashboard. QUE Batch Import accepts JSON files with a single array of objects, Excel files with a header row, and CSV or TSV files. The guide on handling webhook callbacks from a render service covers the callback side. The official QUE API docs are at https://que-api-docs.dataclay.com/, and QUE webhooks are documented at support.dataclay.com/que/content/que/integrations/que_webhooks.htm.
Finally, test with the same tools you use for any batch. Templater’s Spot Checking lets you quality check a batch render by seeing how data merges across jobs, and Dataclay recommends a pre flight render of one video. If your pipeline pulls data for a live event, the Live Sports Stats Autograf is built to link data with Google Sheets or JSON files.
Pipeline Design
Five choices before you go live with a feed
-
Update render-statusBecause The Bot cannot write it back to a URL feed, register a script or have your application do it, so that ready data is not rendered again.
-
Give every row a stable labelDataclay recommends a column named id to label each spreadsheet row for Templater output, and the same idea applies to a JSON property key named id for a feed.
-
Match the keysProperty keys in a URL feed must match dynamic layer names or reserved keys, so check your endpoint’s field names against your layer names.
-
Think about the machinesDataclay recommends disabling After Effects disk caching on any machine where Templater Bot is left enabled.
-
Test with a small batchTemplater’s Spot Checking lets you quality check a batch render by seeing how data merges across jobs.
The QUE Route for Job Data
If you send jobs from your own application, QUE is the other path the documentation describes. Templater and QUE use render-status to show a job’s progress through the pipeline, and The Bot processes QUE jobs that have a render-status of ready. A new QUE job has a render-status of ready by default, and the QUE API default render-status values are ready, queued, processing and done. In the QUE API, jobStatusFilter set to true filters jobs by the render-status given in jobStatus.
A QUE job can be inserted into a campaign with the Create Job API endpoint or from the QUE Dashboard. A new job can be made with a File Based Method (Select JSON File) or a Manual Entry Method, and when a job is created from a JSON file, the file’s values override values from the campaign’s defaults. To use Google Sheets with QUE, set QUE as the Templater data source and connect the Sheet via the QUE API or Zapier.
Bottom Line
The documentation does not give an exact moment when a live API feed becomes a liability, so that question stays open. It does tell you where a feed needs attention. The Bot cannot update render-status on a URL feed, so you must add a script or application step to keep ready rows from being rendered again. The Bot reads a feed no more often than every 12 seconds, queues up to 20 rows at a time, and relies on property keys that match your layer names. Error codes 6001, 1002 and 1003 name the HTTP and footage download failures. Plan for those facts, run a small test, and start from Templater for Adobe After Effects for the product overview.
Frequently Asked Questions
When does a live API feed become a liability in a render pipeline?
Dataclay’s documentation does not name an exact moment. It does document the limits of a URL feed: The Bot cannot update the render-status key, so ready data can be rendered again unless a script updates it. Test your own feed with a small batch to see where it needs extra care.
What is a Templater URL feed?
A Templater URL feed is a REST API endpoint that returns responses as JSON object arrays. You link it by entering its address in the Data source field in Batch Assets.
Can The Bot update render-status on a URL feed?
No. With a URL feed, The Bot cannot update the render-status key, so users should register scripts to update it. Dataclay advises custom scripts so that ready data is not rendered again.
How often does The Bot read a URL feed?
The default polling interval is 0.5 minutes, or 30 seconds. For URL feeds and JSON sources, the minimum interval is 0.2 minutes, about 12 seconds.
Which error code means a failed HTTP request?
Templater error code 6001 means HTTP request error. Code 1002 means error downloading a remote footage source.
Does setting up a URL feed need technical skills?
Yes. Setting up a URL feed as a Templater data source requires knowledge of configuring web servers.
How many rows can The Bot queue at one time?
Templater Bot can queue up to 20 rows or JSON objects at one time, which is both the default and the maximum.
