With the tremendous increase in multimedia content on the Web and in corporate intranets, discovering hidden meaning in raw multimedia is becoming one of the biggest challenges. Analysing multimedia content is still in its infancy, requires expert knowledge, and the few available products are associated with excessive price tags, while still not delivering sufficient quality for many tasks. This makes it almost impossible for normal companies, particularly SMEs, to make use of this technology. Also, analysis components typically operate in isolation and do not consider the context (e.g. embedding text) of a media resource. MICO addresses these problems as follows:- it will consider "media in cross-media context", allowing to analyse media resources as well as connected content, including video, images, audio, text, link structure and metadata- it will investigate cross-media analysis along the complete, distributed analysis chain, namely extraction, metadata publishing, querying and recommendations- it will develop harmonized models and software services for orchestration of analysis components, representing and publishing of analysis results, and querying and evaluating such results- it will contribute its main software development results as Open Source components to two established Apache projects, Apache Marmotta and Apache Stanbol, simplifying the use of the technology in industrial productsThe technology will be evaluated in two complementary use cases: crowd sourcing for academic research at Zooniverse (most successful crowd science platform) and video sharing at KIT digital (one of the biggest enterprise video sharing providers). Through its involvement in the Apache Software Foundation, the project partners will reach out to many technology companies and have access to a large community already at the beginning of the project. Publishing the main results as Open Source will enable particularly SMEs to benefit from cross-media analysis technology.