CUSTOMSCREENCOMPUTER-USE
Software Task Screenflows
Screen workflow recordings with task objectives, OS/app metadata, cursor activity, step boundaries, completion status, and action logs for computer-use model evaluation.
CUSTOMSCREENERROR RECOVERY
Error Recovery and Correction Screenflows
Screen workflow collection format focused on mistakes, failed attempts, correction steps, undo sequences, retries, and recovery behavior inside approved software or browser tasks.
Error Recovery and Correction Screenflows are built for teams evaluating how agents behave when workflows do not go perfectly. The format captures mistakes, wrong inputs, failed navigation, validation errors, undo sequences, retries, corrections, and eventual completion or abandonment.
Each session is structured around a task where recovery behavior is expected, induced, or naturally observed. Screen recordings can be paired with error event markers, correction-step annotations, cursor or keyboard events, task metadata, recovery strategy labels, QA fields, and delivery records.
Structured collection
Error Recovery and Correction Screenflows are structured around tasks where the contributor encounters, creates, or resolves an…
Metadata depth
The format supports alignment between screen recording timestamps, error markers, action logs, workflow phase labels, and…
Delivery-ready package
Includes screen clarity, error marker, correction sequence reviews, plus dataset manifest, CSV catalog, consent status metadata,…
Request sample access
Share your target workflow, error categories, correction markers, and evaluation goals. We’ll reply with a technical scope for the collection.
Key highlights
- Screenflows centered on error states, correction paths, retry behavior, and workflow recovery.
- Error event markers aligned to timestamps and task progress.
- Optional action logs for undo, redo, retry, correction, failed navigation, and validation events.
Dataset contents
Screen capture and logs
Screen recording files · Task match status fields · Error event markers · Correction step records · Undo and redo event markers
Task and workflow records
Recovery strategy labels · Optional click/key event logs · Task objective metadata · Retry pattern metadata · Completion status
App and session metadata
App or browser metadata · Operating system metadata · Optional cursor coordinates
Technical specifications
Use cases
- Agentic AI robustness testing
- Error recovery policy evaluation
- Computer-use agent debugging
- Workflow failure analysis
- Correction behavior modeling
- GUI automation evaluation
- Task completion under uncertainty
Dataset workflow
1Collection
Error Recovery and Correction Screenflows are structured around tasks where the contributor encounters, creates, or resolves an error condition. The goal is to preserve the sequence from failed action to correction path and final outcome.
2Review
Screen clarity review · Error marker review · Correction sequence review · Task match review · Action log consistency review · Consent status review · Delivery approval
3Delivery
MP4 screen recordings · Error event JSON · Action log JSON where scoped · Workflow metadata CSV · Dataset manifest JSON · CSV file catalog · QA summary · Consent status metadata · Delivery notes
Quality assurance
Related dataset formats
CUSTOMSCREENWEB WORKFLOWS
Browser Navigation and Web Task Flows
Browser task recordings with URL context, click and scroll events, form interaction metadata, page categories, task outcomes, and action logs for web-agent evaluation.
CUSTOMSCREENMOBILE WORKFLOWS
Mobile App Task Screenflows
Mobile screen recordings with tap events, swipe events, app categories, task objectives, step boundaries, completion status, and device metadata for mobile-agent evaluation.
CUSTOMSCREENNARRATED WORKFLOWS
Guided Software Walkthrough Sessions
Narrated software walkthroughs with screen recordings, aligned narration transcripts, step markers, app context, cursor flow, and task metadata for procedural software understanding.
CUSTOMMULTIMODALPAIRED DATA
Multimodal Paired Capture Sets
Paired multimodal datasets linking video, audio, screen, text, or image-derived records through shared file IDs, cross-modal manifests, metadata schemas, and QA fields.
CUSTOMANY MODALITYBUYER-SCOPED
Buyer-Scoped Custom Collection
Fully scoped custom data collection with buyer-defined modality, schema, metadata fields, annotation layers, QA criteria, manifest structure, delivery package, and collection rules.