We build AI features that take work off the plate of newsrooms and broadcasters. They are based on large language models such as GPT, integrated via LangChain, RAG and vector stores, with components in Python, TypeScript, Rust and Go.

In the FC Augsburg app, AI suggests photo captions and push notifications to the editorial team. For radio stations our systems sort songs into playlists automatically, and we analyze Germany's largest music playlists.

What we deliver:

• GPT applications for chat, search and automation

• Automatic classification of music, podcasts and texts

• AI assistance for editorial teams, for example for copy and push notifications

• Integration of LLMs into existing systems and workflows

• Agents with memory and tool use based on LangChain

• Connections to databases, APIs and CMS via RAG and vector stores

• Scalable infrastructure with Rust, Go and Docker on AWS or Hetzner