Version 1.1.61
This update marks a major leap forward in performance, reliability, and flexibility for TaskBots. It’s one of our biggest and most exciting releases yet! Here are some of the major highlights:
Overhauled data processing and TaskBot architecture for faster, more reliable runs, especially on large workflows.
Take screenshots directly from your TaskBots.
Navigate directly to any building block from logs, error reports, or run dialogs.
Set a custom starting building block with the new "Start run from here" option.
Highly robust scrolling behavior even on complex, virtualized lists.
Improved anti-bot detection prevention on actions for clicks and scrolls.
More flexible and powerful looping, including custom iteration tracking.
Custom error handling is now possible with Try-Catch and Raise Error building blocks.
Else condition added to condition blocks for smoother conditional flows.
...and so much more! Dive into the full notes below. 👇
🔥 New Building Blocks
📸 Take Screenshot Capture full pages, specific web elements, or visible areas and save them locally. The saved path can be added to variables or table columns, making it easy to upload screenshots later with the Upload File building block, if needed.
🔀 Try-Catch, After Try-Catch and On Catch Error Wrap any building block or sequence of building blocks inside a Try-Catch. If an error occurs, the run will continue along a separate failure path (On Catch Error), letting you define exactly how to handle errors.
⚠️ Raise Error Raise your own custom errors within a TaskBot run. Perfect for generating custom error reports or triggering custom error handling flows through Try-Catch blocks.
✨ New Options in Existing Building Blocks
Set Condition: New Else condition path added, along with "is a valid number" and "is not a valid number".
Number Operations: Added Remainder % operation.
Save Web Element: New option Skip if no element is found, allowing the run to continue without errors if elements are occasionally missing, especially useful when the Auto-scroll option is enabled in loops.
Log: Now supports tag and type (success, warning, fail) for more organized logging.
Split Data: Now supports negative numbers in position fields to split from the end of text.
Ask ChatGPT: Added support for the newly added models.
⚡ Improved TaskBot Run Behavior
Improved Anti-Bot Detection: Improved actions like click and scroll to bypass stricter anti-bot detections on sites like social media platforms.
File Handling: The Save File building block now waits for download completion before the TaskBot run is continued. There is now also auto-retry in place when a file download from a URL times out.
More Robust Save Web Element: There is now additional handling in place in case the web element is found but its content was not yet loaded by a website.
Spintax Limit Increased: From 100 to 5,000 expressions thanks to optimized handling.
Improved Send HTTP Request Building Block:
Added auto-retry on 429 (rate limit) and network timeouts.
When no matching columns are found between the response keys and the target table, no data is saved instead of throwing an error.
🔁 Looping Improvements
Custom Continuation: The Start Repeat building block now supports a custom "Start from" input.
Improved Continue at Latest Iteration: More reliable tracking of the last row or iteration.
Auto-Scroll Overhaul: Highly robust and human-like scrolling, even in quirky or virtualized lists. Auto-scrolling is now also supported when the Continue until no element is found option is enabled.
Loop Index Flexibility:
Use multiple
{loop_index}inside a selector. (Prior to 1.1.61, only one was accepted per selector.)Add loop IDs like
{loop_index_123}when accessing loop index of a parent loop. Perfect for looping through web tables, where the parent loop goes through rows and the child loop through columns — with selectors able to access both parent and child loop indices.The starting point in the
{loop_index}syntax is now optional.
🚀 Performance and Architecture
New TaskBot Run Architecture: TaskBot run flow and all building blocks have been overhauled for faster and more reliable processing.
Data Processing Overhaul:
TaskBots now handle reading and writing dynamically based on the amount of data, reducing the risk of timeouts when handling large amounts of data.
Future integrations (like Airtable and custom SQL storage) are now easier and faster to add.
Faster TaskBot Building: Adding new building blocks to the canvas is now 10x faster. Moving blocks, deleting selected blocks, and adding connections between blocks is faster, too.
Stable Stop Logic: Race conditions between Run and Stop that in rare cases could cause duplicated data are now fully eliminated.
Database Upgrade: Larger, faster database to handle growing usage.
Export Improvements: Faster log exports.
🛠️ Better Building and Troubleshooting Experience
Start Run from Here: Right-click any building block and run starting from there — no need to deactivate anything!
View Building Block: Navigate directly to the building block from logs, errors, or notifications. Works even if another TaskBot is open.
Log Enhancements:
Cleaner, categorized logs with titles, categories, and visually separated building block IDs.
Improved error reports with a more readable format and related documentation links.
Note: These improvements apply only to new logs and error reports; the existing ones remain unchanged.
Better Commenting:
Save TaskBot descriptions and building block comments with
CMD+S/CTRL+S.The editing experience is now organized in a smoother flow.
Better Search in Logs: Improved search functionality inside TaskBot run logs.
Unique TaskBot Names: Duplicate names no longer allowed to avoid confusion.
Persistent Input on Validation Errors: Inputs (like building block comments, TaskBot descriptions, TaskBot names, and variable names) are preserved if a validation error occurs.
Improved Deactivation Handling: Smarter path selection if conditions or paths are deactivated.
Smoother Duplicating: When duplicating building blocks, it is now more seamless to move and position the newly created blocks.
Warning Tag for Unsaved Changes: When you have unsaved changes inside a building block, a tag now appears to warn you. It’s non-intrusive — you can still close the dialog without saving if you want — but it gives a clear visual reminder in case you forgot to save.
Better Validation Before Starting a Run: New validations help prevent invalid setups. But also validation is now more flexible: for example, if multiple connections are deactivated (or only one is active), the run is still allowed, making it easier to test multiple paths.
🖌️ Minor UI Improvements
Proper Mac key labels (CMD/Option) in the Keyboard building block.
Cleaner text field scrollbars.
Improved drag-and-drop styling for building blocks.
Send Notification emails now include a standard footer.
TaskBot descriptions support line breaks.
🐛 Bug Fixes
Fixed performance lag caused by text field re-renders due to a bug with autosize calculations.
Fixed an issue where the Abort building block did not properly abort when used inside nested loops.
Fixed sheet reference parsing when column names included line breaks.
Fixed all known unicode-related issues.
Fixed JSON path filtering in webhook dynamic data.
Fixed tab switching issues with the Switch or Close Tab building block.
Fixed incorrect log numbering in TaskBot run reports.
Fixed UI cache sync issues for TaskBot name and description.
Fixed a bug where webhook payload size was limited incorrectly.
🔒 Security Updates
Improved email sanitization for the Send Notification building block.
Strengthened data ownership validation on all endpoints.
Added internal monitoring for suspicious requests.
⚠️ Changed Behavior
"Create file column" moved under "More options" in native tables.
Updated variables are now sent only at the end of the run (not after each loop end). No impact on how TaskBots run.
Error report screenshots now retained only for 14 days.
Log limit reduced to 5,000 log messages per TaskBot run.
It is no longer possible to search for timestamps inside the log in the TaskBot run reports.
Per-minute rate limit added to the Send Notification building block.
Maximum of 10 email addresses allowed per TaskBot for run notifications. Existing setups with more than 10 emails will continue to work — no breaking changes.
Dynamic loops no longer apply implicit filtering. TaskBots now always iterate through all rows in a table. This does not break existing runs (conditions still apply).
"None" input is no longer available in Continue at latest iteration. TaskBots using "none" will keep working — no breaking changes. To skip the loop, simply set the repetition limit or number of repetitions to 0.
Character limit added to messages in
await log()inside the Write JS building block. If the logged message exceeds the limit, it will still be logged but in truncated format.Custom error handling removed from Ask ChatGPT. TaskBots that used custom handling will still work — no breaking changes — but it’s recommended to move error handling to Try-Catch building blocks.
"Stay on page after run" option now respected for scheduled and webhook runs. If toggled on, the browser will stay open after the run finishes. (Previously, the setting was ignored if it was a scheduled or webhook run.) Be aware that browsers could now remain idle if this option is unintentionally left enabled.
🚨 Potential Breaking Changes
Save Page URL and Save from Clipboard building blocks no longer have built-in delay options. If you relied on the old delay behavior, you may need to add a separate Delay block.
In Start Repeat, the option "Auto-detect the number of available elements" is now renamed to "Continue until no element is found," and the lead selector is no longer required. For most TaskBots, nothing will break — but if you relied on a specific lead selector (especially when selectors inside the loop are unrelated to it), you could end up with more or fewer loop iterations than before.
In Save Web Element, the "Stop if no element is found" option is replaced by "Skip if no element is found." Existing TaskBots will not break — the new behavior matches the previous setting automatically — but if you were relying on it to force early termination, you may need to adjust your logic. Otherwise, it could now result in an infinite loop if no elements are found.
Auto-scroll is now triggered by default in all loop setups (except dynamic loops that iterate over tables). Scrolling behavior has been significantly improved, but if you previously implemented custom scrolling (for example, with a Write Javascript building block), you should double-check your setup. You can disable Auto-scroll in the Start Repeat building block if needed.
In Set Condition, "Longer than" and "Shorter than" now only accept number of characters as input. Previously, it was possible to compare the character count of two texts.
Input validation during TaskBot runs is now stricter. Previously, invalid inputs could potentially pass silently. If needed, you can add Try-Catch blocks to handle these cases gracefully.
Last updated
Was this helpful?