cursor
Datasets
All datasets matching “cursor”cursor-deliveries-publicFable-5-Cursor-TracesFable 5 Cursor Traces
244 Fable 5 Cursor agent sessions for training & research.
This dataset has 244 Cursor sessions with Fable 5 at High/xHigh/Max effort levels for distillation.
[!IMPORTANT]
This dataset is compatible with Teich! Use it directly in your Teich training pipeline.
[!WARNING]
The longest rows exceed one million characters of content. Apply prepare_data() with your intended tokenizer and an explicit context/oversize policy before training.… See the full description on the dataset page: https://huggingface.co/datasets/TeichAI/Fable-5-Cursor-Traces.cursor-traces-exampleThis dataset was generated using teich by TeichAI
My Agent Traces
This directory contains raw agent trace files generated by teich.
JSONL files: 9
Training-ready tools
Generated agent traces carry configured or recovered tool schemas so tools remain available for training even when a session did not call them.
Native Claude Code imports recover schemas for Claude Code and Claude Desktop built-ins, plus conservative name-derived MCP schemas, when the raw… See the full description on the dataset page: https://huggingface.co/datasets/armand0e/cursor-traces-example.cursor-groundingcursors
Cursors
A dataset of labelled OS cursor images for compositing onto cursor-free screenshots as synthetic training data for GUI grounding models.
366 rows across 7 OS/theme sources, covering 31 cursor types.
Intended use
Given a screenshot with a cursor composited into it, a model is asked:
"What cursor type is shown in the red box?"
The cursor_type labels are chosen to be visually distinguishable — if two themes render a cursor name differently (e.g. one theme draws move… See the full description on the dataset page: https://huggingface.co/datasets/Fraser/cursors.cursor-streams
Cursor Streams
The largest open human computer-use video dataset combining precise mouse-event annotations with frame-by-frame cursor localization.
Cursor Streams is a 156-hour human computer-use dataset containing 11,746 desktop interaction trajectories, 16,848,718 frame-level cursor annotations, and 61,804 precisely timestamped mouse interaction events across 105 foreground applications.
It combines continuous screen video with accurate click timing and dense frame-by-frame… See the full description on the dataset page: https://huggingface.co/datasets/Cianmcnally/cursor-streams.