getting started with managed agents in the gemini api: an inside look with Logan Kilpatrick, Philipp Schmid, and Ali Çevik

@OfficialLoganK @alihcevik Great conversation! If you have some questions, complains or ask please let me know.

@OfficialLoganK @_philschmid @alihcevik رحلة ممتعة في أعماق التقنية يا صديقي! استكشاف الوكلاء المدارين (Managed Agents) عبر Gemini API هو خطوة جوهرية نحو فهم مستقبل التفاعل الذكي وبناء أنظمة أكثر استقلالية. شكراً لهذا الإثراء المعرفي. ✨

@OfficialLoganK @_philschmid @alihcevik Keep your agents secure!

@OfficialLoganK @_philschmid @alihcevik You're making some cool agents. Respect.

@OfficialLoganK @_philschmid @alihcevik the real unlock isn't just using api, it's figuring out where the edges are

@OfficialLoganK @_philschmid @alihcevik Nice, always cool to see how people are actually using this stuff in production. Curious what surprised you most while building with managed agents.

@OfficialLoganK @_philschmid @alihcevik

@OfficialLoganK @_philschmid @alihcevik You can share my English tense aspect I do not want any favor. Free of charge the memory code model is in the article

@OfficialLoganK @_philschmid @alihcevik another day closer to agents doing my entire job while i nap.

@OfficialLoganK @_philschmid @alihcevik Thank you for these short videos. A great idea IMO. Useful in that it provides a high level overview, examples and direction for further exploration / education for me.

@OfficialLoganK @_philschmid @alihcevik The model scoreboard changes fast. The durable edge is workflow fit: which system handles tools, memory, and weird failure states best.

@OfficialLoganK @_philschmid @alihcevik Здравствуйте. Ребята, вы такие бравые, формы подтянутые и шеи прям как у спецназа...когда вы кодом занимаетесь ? Вы точно программисты? )) Шучу... Видно сразу мастеров автоматизации. Спасибо за настройки и обновления.

@OfficialLoganK @_philschmid @alihcevik Managed agents are useful when teams need a clearer path from prototype to production. The practical test is how well state, tool calls, permissions, and failures can be observed and controlled.

@OfficialLoganK @_philschmid @alihcevik Great!

@OfficialLoganK @_philschmid @alihcevik the managed agents setup looks clean, curious how the memory/context works across sessions. been building nearby and this could simplify a lot

@OfficialLoganK @_philschmid @alihcevik Curious how these managed agents handle unexpected user input or requests outside their defined scope. Does it fail gracefully or try to improvise?