Graviti: Delivers production-grade annotated real-world datasets for physical AI, but lacks general text or audio annotation workflows. Defined.ai 每月有 146,251 次造訪,月增率 -21%,定價模式為需聯絡報價。Graviti 每月有 755 次造訪,成長率 -82.8%,定價模式為免費增值。
| 面向/指標 | ![]() | |
|---|---|---|
| 官方網域 | definedcrowd.com | graviti.com |
| 定價模式 | 需聯絡報價 | 免費增值 |
| 每月造訪次數 | 14.63萬 (146,251) | 755 (755) |
| 月流量變化量 | -3.88萬 (-21%) | -3622 (-82.8%) |
| 主要市場與市佔 | 🇺🇸 美國 (20.6%) | 🇮🇳 印度 (100%) |
| 每月直接造訪 | 8.44萬 | 290 |
| 使用者評分 | 4.4 | 3.8 |
| 支援平台 | 網站 | 網站 |
企業級 AI 訓練資料、標註與評估服務
提供涵蓋音訊、圖像、影片、文字與多模態格式的大量合規來源資料集,支援進階篩選與免費樣本。
面向文字、音訊、圖像與影片的專家驗證標註,採用模型優先、人在迴路的工作流程,兼顧準確性與公平性。
面向全球 AI 系統的高品質多語言內容,覆蓋 500 多種語言、方言與語區。
面向訓練的大規模全球化多元資料蒐集,每個資料集都經過隱私與安全審查。

面向具身智慧、機器人與大模型的真實世界訓練資料
旗艦資料集將頭戴式攝影機與多台固定攝影機配對,同步錄製同一場景並完成校正,讓模型學習空間化的動作理解。
為機器人操作與世界模型研究擷取並標註雙臂遙操作軌跡、人體關鍵點、動作分段與語言指令。
面向語言模型,Graviti 提供專家撰寫的 STEM 題目、用於 RLHF 與 DPO 的人類偏好配對、代理程式編程環境,以及專家評分的金融、法律與辦公任務。
訓練有素的標註員補充動作分段、手部與人體關鍵點、敘述、偏好與理由,隨後經過多輪人工審核、自動化檢查與污染篩查。
免費 / 開源ModelScope (魔搭社区) is an open-source AI community and model service platform that hosts thousands of pretrained models, datasets, and ready-to-run demos, backed by Alibaba Cloud infrastructure.
免費 / 開源ModelScope (魔搭社区) is an open-source AI community and model service platform that hosts thousands of pretrained models, datasets, and ready-to-run demos, backed by Alibaba Cloud infrastructure.
免費 / 開源簡化了電腦視覺資料集標註與模型部署,但缺乏 Defined.ai 廣泛的人工群眾外包服務。

SuperAnnotate is an AI data platform that combines a collaborative multimodal annotation editor, an expert data-operations team, and the orchestration layer needed to move curated datasets into training, evaluation, and production pipelines. It is aimed at teams building computer vision, NLP, audio, and generative AI systems who need reliable, high-quality ground truth at scale.
SuperAnnotate is an AI data platform that combines a collaborative multimodal annotation editor, an expert data-operations team, and the orchestration layer needed to move curated datasets into training, evaluation, and production pipelines. It is aimed at teams building computer vision, NLP, audio, and generative AI systems who need reliable, high-quality ground truth at scale.

面向具身智慧與機器人訓練的影片資料基礎設施

免費 / 開源Label Studio is an open-source data labeling and AI evaluation platform from HumanSignal. It handles images, video, audio, text, documents, time series and agent traces in one flexible workspace, and connects to your ML pipeline through an API, Python SDK and webhooks.
免費 / 開源Label Studio is an open-source data labeling and AI evaluation platform from HumanSignal. It handles images, video, audio, text, documents, time series and agent traces in one flexible workspace, and connects to your ML pipeline through an API, Python SDK and webhooks.

免費 / 開源Voxel51 is the data platform for physical AI, built on the open-source FiftyOne library. It helps computer vision and robotics teams search, visualize, curate, annotate, and evaluate multimodal datasets — images, video, 3D scenes, and sensor streams — so they can find where models fail and fix the data behind them.
免費 / 開源Voxel51 is the data platform for physical AI, built on the open-source FiftyOne library. It helps computer vision and robotics teams search, visualize, curate, annotate, and evaluate multimodal datasets — images, video, 3D scenes, and sensor streams — so they can find where models fail and fix the data behind them.

Provides expert human workforce for AI training and evaluation datasets, though with fewer automated labeling tools.
Provides expert human workforce for AI training and evaluation datasets, though with fewer automated labeling tools.