Trickest Changelog

​

Trickest Major Update

​

Global redesign changes

• Navigation - Unified left sidebar with a workspace switcher and global search (⌘K). Navigation is now more consistent across the app (same placement, same icons), so you can move between Workflows, Library, Files, Variables, and other areas without context switching.
• List pages - Standardized list layout across pages (header, search, filters/sort, results). Lists emphasize scanability: clear columns/metadata, consistent status indicators, and always-visible result counts. Common actions are available via row-level ... menus.
• Cards - Cleaner, consistent card design with a predictable metadata order (name, key stats, timestamps/status). Cards use the same interaction pattern across entities (hover/select + ... menu), reducing one-off UI behaviors.
• Empty states - Actionable empty views with a clear primary CTA (for example, create a workflow, upload a file, or add a variable), plus short guidance on what to do next. This reduces “dead ends” when a workspace/project is new.
• Detail panels - Slide-in detail panels for inspecting objects (for example, runs, library items, and profiles) while keeping your current page and selection intact. You can review details and close the panel to continue where you left off.
• Visual consistency - Unified theme, spacing, and typography; standardized button styles and iconography. Loading and feedback patterns are now consistent (skeletons/spinners where applicable, plus reliable toast notifications for success/failure).
• Information hierarchy - Clearer page titles/descriptions and improved use of badges for quick scanning (status, Partial, etc.). Content is grouped more logically (for example, runs by recency, variables by scope, workflows by project) to reduce visual noise.

​

Workflow building, debugging and execution

• Nodes I/O system - Inputs are now configured directly in each node’s run window (no explicit input nodes).
• Nodes connection system - Connections between nodes are simplified to a single mapped line: when you connect nodes, you choose which upstream output to bind to each downstream parameter. Open the downstream node to inspect the resolved input content and validate the mapping before you run.
• Single-node run and Memoization - You can now execute a single node inside a workflow or resume execution from any specific point - without re-running the entire workflow. Outputs from previous jobs are cached automatically, so earlier steps are treated as if they just ran. This makes iterating on workflows significantly faster and reduces unnecessary compute usage when only part of a workflow needs to be re-executed.
• Job distribution (splitter) - Distribution is now built into the node run experience. Previously, you had to add and wire an explicit splitter node (and helpers) to fan out work across machines. In the new editor, any node that supports parallel execution (based on its function and the shape of its input) can be distributed directly from its run window. By enabling Distribute, the platform splits the incoming input into parallel jobs and runs them concurrently. You can distribute by file lines (one job per line), by folder (one job per file), or by file batches (Files in groups)—where a file is split into at most as many batches as machines, so job scale matches your run. This brings several benefits like running large-scale jobs with minimal setup, faster iteration, and better per-job visibility.
• Builder and Runs tabs -> Workflow tab - The workflow editor is now unified under a single Workflow tab. Previously, you built workflows in Builder and reviewed executions in a separate Runs tab. Now, design, execution, debugging, and results are all available in one place: you can run workflows, browse run history, inspect node inputs/outputs, and restore prior versions without changing the context.
• Building modules - Modules are built on their own dedicated editor page, just like regular workflows. The Module I/O section automatically detects all inputs and outputs from your workflow nodes and lists them as candidates for the module’s interface. You choose which to expose, preview output values before confirming, and rename or remove any entry at any time. The module saves automatically and remains immediately available in the Library and the editor’s node picker.
• Undo and Redo - You can undo and redo changes on the canvas (e.g. adding, deleting, or moving nodes) using Cmd+Z / Cmd+Shift+Z (Mac) or Ctrl+Z / Ctrl+Shift+Z (Windows/Linux), or the undo/redo controls in the bottom-left vertical toolbar on the canvas.

​

Automatic Data Format Detection for Live Tables