AgentNet
22.6K human-annotated desktop computer-use tasks on Windows, macOS and Ubuntu, released with OpenCUA for training and evaluating vision-language-action models.
Release date not recorded · desktopBrowse 86 reviewed datasets connected to computer-use, GUI, web, and mobile agents. Each entry links to primary evidence and records its scope.
Review means AI-assisted editorial inspection, not independent reproduction. Unknown licenses and specifications remain unknown.
Compare source-reviewed specifications
22.6K human-annotated desktop computer-use tasks on Windows, macOS and Ubuntu, released with OpenCUA for training and evaluating vision-language-action models.
Release date not recorded · desktopPublished AgentSynth task definitions and recorded GUI trajectories, with OSWorld and InSTA training configurations and packaged screenshots, thoughts and commands.
Release date not recorded · desktop, webSynthesizes browser-agent trajectories by replaying web tutorials as guidance, releasing the trajectory data with multimodal grounding and reasoning and an AgentTrek-32B agent.
Release date not recorded · webThe first-stage Aguvis training collection, packaging GUI element-grounding annotations and image archives from multiple interface-data sources.
Release date not recorded · cross-platformThe second-stage Aguvis collection of interface trajectories for learning action sequences, planning and reasoning after grounding pretraining.
Release date not recorded · cross-platformCode for AMEX Android GUI data that converts its three annotation levels into question-answer pairs and fine-tunes a LLaMA2-Accessory model, with a released checkpoint.
Release date not recorded · androidHuman demonstrations of Android tasks with screenshots and touch actions.
Release date not recorded · androidHTML and annotation release used for Code2World’s action-conditioned GUI-state prediction training, constructed from mobile GUI trajectories.
Release date not recorded · androidAndroid demonstrations with high-level goals, step instructions, screenshots, and accessibility trees.
Release date not recorded · androidEpisode-based mobile GUI grounding data released with Aria-UI for incorporating interaction history into target localization.
Release date not recorded · androidGUI element descriptions and grounding instructions spanning web, mobile, and Ubuntu desktop interfaces.
Release date not recorded · web, android, desktop, linuxConversation-formatted supervision released for fine-tuning AutoDroid V2, with the project documenting a ShareGPT-style JSON configuration.
Release date not recorded · androidScreenshot-based grounding and captioning samples with UI functionality descriptions, target coordinates, and bounding boxes.
Release date not recorded · web, cross-platformChinese Android GUI benchmark released with AgentCPM-GUI, with a grounding track for widgets and OCR text and an agent track for multi-step tasks in Chinese apps.
Release date not recorded · android18k trajectories with 98k steps from more than 50 Chinese mobile apps in domains such as e-commerce and social media, for building GUI agents on Chinese apps.
Release date not recorded · androidComputer-use environment and task resources for agent research.
Release date not recorded · desktopDrag-grounding training and evaluation data pairing each screenshot instruction with start and end boxes, across text highlighting, spreadsheet cells, slide resizing and sliders.
Release date not recorded · cross-platformWeb-world-model training data with prompts, images, action descriptions, and responses, released with the WebDreamer project.
Release date not recorded · webInstruction-to-function-call examples for learning Android intent invocation, with training, testing and API-annotation files for mobile task execution.
Release date not recorded · androidOver 20,000 human Android episodes across 500+ apps with user context and long-term usage patterns, plus tracks for proactive intent prediction and personalized execution.
Release date not recorded · androidFlattened web-interaction training records with accessibility-tree observations, assistant completions, trajectory rewards and step-link metadata.
Release date not recorded · webDesktop grounding prompts selected from GroundCUA and formatted for click-based reinforcement learning with target-box rewards.
Release date not recorded · desktopThe public training archive linked by the Phi-Ground project for its Phi-Ground-Any model.
Release date not recorded · unspecifiedDesktop GUI trajectory data supporting multiple interface understanding tasks.
Release date not recorded · desktop, windowsAction narration resources for GUI interaction trajectories.
Release date not recorded · cross-platformA processed screenshot and bounding-box collection released for training GUI-Actor visual grounding models.
Release date not recorded · unspecifiedA published JSONL corpus of synthesized planning and causal knowledge for GUI-agent mid-training, linked directly from the official GUI-CIDER project.
Release date not recorded · androidData for learning text-dragging interactions in graphical interfaces.
Release date not recorded · cross-platformThe published GUI-Lasagne training-data collection used for SpiritSight, distributed as dated archive shards for visual interface learning.
Release date not recorded · unspecifiedA TongUI-collected training corpus from multimodal web tutorials, distributed as annotation files and split screenshot archives for GUI trajectory supervision.
Release date not recorded · unspecifiedHuman-demonstrated Android episodes for cross-app navigation, spanning six task types, 212 apps and about 1,400 app combinations, for training and evaluating agents.
Release date not recorded · androidTraining and evaluation Parquet bundles accompanying GUI-R1, including GUI grounding and mobile/web action-prediction data.
Release date not recorded · desktop, windows, linux, macos, android, webGUI video benchmark and training data covering six scenarios, from Android and iOS to websites and XR, with eight question types for image and video LLMs and the GUI-Vid model.
Release date not recorded · cross-platformWeb and Android GUI action-instruction data, with single-step, multi-step and smartphone partitions.
Release date not recorded · web, androidConversational GUI supervision with text-rich images and bounding boxes, organized as single-turn questions and multi-turn dialogues.
Release date not recorded · unspecifiedThree datasets for turning general VLMs into GUI agents: GUIEnv for OCR and grounding, GUIAct with 67k single-step and 15k multi-step web and Android actions, and GUIChat dialogues.
Release date not recorded · cross-platformThe released local-supervision portion of GUIEnv, pairing webpage regions and text for OCR and GUI grounding.
Release date not recorded · webA released GUI-agent mid-training mixture and supporting domain datasets, together with a post-training trajectory file and image archives for web and mobile interaction.
Release date not recorded · web, androidCross-app Android interaction episodes with fine-grained step instructions, screen descriptions, action context, and target boxes.
Release date not recorded · androidSynthetic GUI grounding data associated with UI decomposition research and the OSWorld-G benchmark.
Release date not recorded · desktopPre-training corpus for Xiaomi's MobileVLM, distributed as per-app graph archives for mobile apps such as Baidu Map, QQ Music, Ctrip and Vipshop.
Release date not recorded · androidOver 600,000 Android screenshot and view-hierarchy pairs from more than 20,000 Google Play apps, collected with an optimized DroidBot, with complete per-app traces.
Release date not recorded · androidCross-platform mobile GUI trajectory dataset derived from instructional videos, with released data-processing helpers and example evaluation code for Android and iOS data.
Release date not recorded · android, ios, cross-platformVideo-derived mobile navigation annotations that pair scene boundaries and screen or element boxes with action labels for training and evaluating GUI agents.
Release date not recorded · android, iosMobile task demonstrations and natural-language instructions.
Release date not recorded · androidMulti-turn web-navigation conversations constructed from Mind2Web interactions, with dialogue turns, HTML snapshots, element candidates and click/type/select action sequences.
Release date not recorded · web7.3 million training samples built from webpage UIs and structured with enhanced accessibility trees, for text-rich visual understanding, released with UIX-Qwen2 models.
Release date not recorded · webBrowser-interaction data released with NNetNav for its live-website setting, with separate training and test files.
Release date not recorded · webNNetNav's WebArena-oriented browser-interaction data release, with separate training and test files.
Release date not recorded · webMobile-agent supervision from emulator and simulator trajectories, plus executable tasks with starting states and success checks for reinforcement learning.
Release date not recorded · android, webGUI grounding pre-training data for OS-Atlas in desktop (Linux, macOS, Windows), mobile and web domains, with element boxes stored as relative left, top, right and bottom ratios.
Release date not recorded · cross-platformInteraction-driven pipeline that synthesizes mobile and web GUI agent trajectories without human supervision or predefined tasks, via reverse task synthesis and a trajectory reward model.
Release date not recorded · cross-platformReverse-task-synthesis mobile trajectories released in separate AndroidControl and AndroidWorld training annotation files.
Release date not recorded · androidReleased dataset used to train PC-Agent-E, linked by the official pipeline for refining human trajectories and synthesizing action reasoning.
Release date not recorded · desktop, windows, linuxPlanner and action-executor training examples for WebArena-Lite and WebVoyager, with model-specific processed files and raw trajectory annotations.
Release date not recorded · webMobile app screenshots, view hierarchies, and interaction traces for UI research.
Release date not recorded · androidAnnounced GUI trajectory synthesis resources for recovery from policy-induced errors; the inspected repository marks code and data releases as in preparation.
Release date not recorded · cross-platformCross-platform training data released with ScaleCUA, with grounding and navigation annotations for Ubuntu, Windows, macOS, Android, iPhone and web screenshots.
Release date not recorded · cross-platformArchives of ScienceBoard agent evaluation runs, separated by agent configuration and screenshot, accessibility-tree or set-of-marks observation mode.
Release date not recorded · desktop, linuxMobile screen descriptions for language-based screen understanding.
Release date not recorded · androidGUI dragging data with screen observations, prior state, and continuous pointer actions, provided as trajectory files and Parquet representations.
Release date not recorded · desktopAbout 86,000 human-written question-answer pairs over 35,000 Rico mobile screenshots, plus ScreenQA Short and ComplexQA variants, for benchmarking screen content understanding.
Release date not recorded · androidGUI screenshots paired with localization instructions and targets, including absent-target examples; image access is gated under publisher terms.
Release date not recorded · desktop, web, android, linuxAnnotation files, task recipes, and image manifests for ScreenRef; screenshots reside in the separately gated image repository.
Release date not recorded · desktop, web, android, linuxHigh-resolution professional interface screenshots and target annotations for evaluating instruction-based GUI localization.
Release date not recorded · desktop, windows, macos, linux7,496 desktop grounding examples built from OmniAct screenshots, with annotations rewritten by GPT-4o to describe element appearance, spatial relations and function.
Release date not recorded · desktopWeb-interface grounding examples with instructions, element boxes and points, paired with a separately packaged screenshot archive.
Release date not recorded · webDrag-and-click episode data for the ShowUI-π model, spanning desktop editing, presentation, handwriting and other interface tasks.
Release date not recorded · desktopGated web grounding training data for UGround in Web-Hybrid and Web-Direct subsets, with a bounding-box version whose coordinates are normalized to 0-999.
Release date not recorded · cross-platformUI-AGILE training annotations and screenshots together with converted GUI-grounding and AndroidControl evaluation files.
Release date not recorded · androidMobile GUI reward annotations released for training UI-Genie’s trajectory-assessment model.
Release date not recorded · androidSynthetic mobile GUI trajectory supervision and screenshots released with the UI-Genie agent-training pipeline.
Release date not recorded · androidTreats grounding instructions as multiple reasoning perspectives, releasing a data pipeline, training code and UI-Ins-7B and 32B models trained by SFT on synthesized instructions, then RL.
Release date not recorded · cross-platformThe released mobile training subset for the original UI-R1 3B model, with action annotations and individual screenshot files.
Release date not recorded · androidA public example subset of the UI-S1 mobile training data, packaged as JSONL annotations and Parquet images.
Release date not recorded · androidDesktop GUI data for interface understanding and interaction research.
Release date not recorded · desktopTwo Rico-derived UI component retrieval datasets: AppSim pairs semantically similar elements across screens, while RefExp links referring expressions to target elements.
Release date not recorded · androidMobile-interface critiques and quality ratings with region coordinates and RICO screen references, including a larger three-annotator release than the original paper dataset.
Release date not recorded · unspecifiedCAD screen recordings paired with action logs and target model images for studying precise interface control and long-horizon visual imitation.
Release date not recorded · webScreenshot-instruction records with target boxes and UI metadata such as OCR, source, element type and platform.
Release date not recorded · unspecifiedTwo JSONL collections for supervised fine-tuning the WebEvolver agent policy and web world model.
Release date not recorded · webWeb-task context, starting observations and action history paired with structured checklists and subgoal analyses for checklist generation.
Release date not recorded · webChecklist-annotated web decision records with page observations, prior actions, one preferred continuation and rejected alternatives for process-reward learning.
Release date not recorded · webWebpage segments, questions, answers, HTML, screenshots and layout metadata for structural reading comprehension.
Release date not recorded · webWeb screenshot and interface data for UI research.
Release date not recorded · webDesktop demonstrations with task descriptions, videos, input logs, and action-linked screenshot evidence for studying reusable agent skills.
Release date not recorded · desktop