Enterprise Knowledge Search
Tianmai Juyuan needed to organize and search a massive broadcast-video archive quickly enough to support timely publishing, short-video creation and intelligent content retrieval.
Reported impact
Tianmai Juyuan gained a combined search-and-production platform that speeds asset discovery, improves tagging quality and makes it much easier to turn live broadcast content into short video quickly.
Tianmai Juyuan used Baidu AI Cloud to transform a giant broadcast archive into a searchable, production-ready media asset platform. The foundation is Baidu's MCA media-content-analysis service, which applies multimodal recognition across video, speech, text, public figures, objects and scenes and then outputs structured tags. Those tags power Yunsou, Tianmai's cloud media-asset search tool, and make it possible to search live and archived content precisely across domestic mainstream media. Baidu also supported a production path beyond search: the micro-editing workflow can ingest live TV streams, automatically clip them, transcode them and publish short videos within minutes. From an architectural standpoint, object storage retains the large video corpus, speech recognition and content analysis extract searchable structure, and the clipping layer turns tagged assets into publishable short-form content. That gives Tianmai a combined archive, search and production stack rather than just a passive content repository.
Broadcast video and live streams are stored in BOS, analyzed by MCA and speech-recognition services to generate structured tags, then surfaced in Yunsou search and fed into the micro-editing workflow for clipping, transcoding and distribution.
Compared with manual tagging, MCA improved analysis efficiency by 3-5 times and structured-tag volume by about 10 times (original: 对比人工视频标注，MCA的分析效率提升了3-5倍，结构化标签数量提升了10倍左右).
MCA analyzes video, speech, text, public figures, objects and scenes and outputs structured tag information (original: MCA能提供音视频内容分析能力。能从视频、语音、文字、公众人物、物体、场景等多个维度进行识别……输出精准的结构化标签信息).
Micro-editing can ingest, clip, transcode and publish new TV content into short videos within 3 minutes (original: 通过微剪能进行实时快速收录和剪辑，自动化拆条、转码，3分钟内就能将新播出的电视节目创作成短视频并发布).
Tianmai Juyuan stores more than 500 channels, 10 years of programs, 20 million hours, 200 million items and 6PB of content (original: 收录全球500多套国内外电视频道，存储了超过10年、2000万小时、2亿多条、6PB的全量节目库和精华片段库).
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Baidu AI Cloud Sample May 15, 2025 CCTV.com CCTV.com needed a better way to surface the value of its huge video archive because title- and keyword-based retrieval could not expose enough scene-level meaning, leaving many useful videos hard to find. Resource retrieval efficiency improved 90%. Multimodal Media Retrieval · Real-Time Speech Recognition Baidu AI Cloud Sample May 15, 2025 The Paper The Paper wanted to use generative AI in news products, but needed a way to answer readers' questions without hallucinations while preserving the accuracy standards of professional journalism. The assistant is grounded on more than 27 million articles and nearly 35 billion characters of original reporting. AI Q&A Assistant · Qianfan AppBuilder · Qianfan ModelBuilder 
Lenovo Sample Oct 13, 2026 Formula E Formula E wanted fan experiences that feel immediate, immersive, personalized, and participatory, but its broadcast and event experience was limited by inconsistent onboard video quality, information spread across… Developed in under 30 days. AI Fan Experience Guide · AI Video Optimization · AR broadcast graphics 