triggerdotdev
Development Feed
Wednesday, Mar 25, 2026
Feature·Merged Mar 25, 2026·M
Scheduled runs route to dedicated node pools
#3271·by myftija
Scheduled runs now prefer a dedicated node pool, reducing contention with on-demand runs that are more sensitive to cold start latency.
Feature·Merged Mar 25, 2026·XL
Model Registry adds catalog pages, comparison dashboard, and AI model tracking
#3270·by ericallam
Users can now browse AI models in a catalog, view detailed metrics per model, and compare performance across multiple models side-by-side. The feature includes a new sidebar section under AI with feature flag controls.
Feature·Merged Mar 25, 2026·XL
AI screen UI gets new icons, timelines, and shortcuts
#3269·by samejr
The AI Prompts and Metrics pages in the webapp now feature custom icons, horizontal timeline displays for span durations, copiable metadata fields, and keyboard shortcuts for filters — significantly improving navigation and usability.
Bugfix·Merged Mar 25, 2026·M
Queue limit errors downgraded to warnings
#3243·by ericallam
Expected validation rejections like queue size limits are now logged as warnings instead of errors, reducing noise in monitoring and alerting systems.
Docs·Merged Mar 25, 2026·M
Import paths standardized and realtime docs reorganized
#3265·by D-K-P
Trigger.dev documentation now includes AI-assisted setup prompts, standardized import paths, and reorganized realtime docs that clearly explain the difference between run status updates and streaming AI output.
Tuesday, Mar 24, 2026
Feature·Merged Mar 24, 2026·M
New MCP tool enables span-level inspection within run traces
#3255·by ericallam
The MCP server now exposes get_span_details, a new tool for drilling into individual spans within a run trace to view timing, properties, events, and AI enrichment data like model, tokens, and cost.
Feature·Merged Mar 24, 2026·
Monday, Mar 23, 2026
Feature·Merged Mar 23, 2026·M
Task runs now track trigger source and action
#3241·by myftija
Each task run now records where it was triggered from — SDK, API, dashboard, CLI, MCP, or schedule — plus the action type (trigger, replay, or test). This enables filtering runs by origin without traversing the run tree and makes it possible to route scheduled runs differently based on their source.
Feature·Merged Mar 23, 2026·
