Kaltura open sources machine-readable AI agent expertise, enabling builders and AI brokers to quickly construct production-grade functions whereas accelerating the shift to agent-first software program ecosystems.
Kaltura has open-sourced a collection of AI agent expertise—structured, production-tested data modules designed for AI coding brokers reminiscent of Claude Code, OpenAI Codex, GitHub Copilot, and Cursor.
In contrast to conventional documentation, these expertise are machine-readable and constantly validated towards manufacturing APIs. They embody examined curl examples, parameter tables, error-handling logic, and safety and compliance patterns, enabling AI brokers to autonomously construct production-grade functions.
As soon as deployed, the abilities remodel AI brokers into immediate platform specialists, decreasing growth timelines from weeks to minutes and enabling end-to-end software lifecycle execution—from construct and configuration to deployment and upkeep.
The capabilities span avatar-driven web sites with real-time multilingual interactions, automated digital occasions configured by way of pure language, interactive AI-powered studying environments, and absolutely automated content material manufacturing pipelines.
Kaltura helps agent discovery and interoperability via 4 requirements: Mannequin Context Protocol (MCP), Agent Expertise Commonplace, llms.txt, and Context7 integration, positioning itself inside an rising AI agent infrastructure layer. That is strengthened by its API-first structure, comprising over 100 REST providers, permitting every functionality to be modularised into AI-consumable expertise.
The providing is complemented by the Unisphere micro-frontend framework, which delivers over 10 embeddable widgets with out construct dependencies or framework lock-in.
“AI brokers are quickly turning into the first means builders and enterprise customers work together with platforms – and most platforms aren’t prepared for that shift. Kaltura’s API-first structure, constructed from day one round programmable providers, is completely fitted to at present’s agentic world. Every ability will get your AI agent from zero to a working, safe, best-practice digital expertise in minutes,” mentioned Zohar Babin.
“By open-sourcing these expertise, we’re inviting all the developer group to construct on our platform – avatar-guided portals, clever content material hubs, AI-powered studying environments – and making each AI agent an immediate Kaltura professional.”
The talents, documentation, and developer assets are actually accessible by way of open repositories, signalling a shift in direction of machine-native, agent-first software program growth.