Hi Hunters,
I'm Patrick, co-founder of the Sumi ecosystem.
We built Kodosumi to solve a problem that we encountered while building agents with CrewAI & Langchain ourselves:
- How do we easily deploy agents, expose an API and make them scale (especially for very long agent runs)?
Since many agent devs are not really infrastructure experts but more domain experts in their respective non-dev fields, we believe providing an easy tool like Kodosumi that handles the scaling part is crucial to enable a whole organisation to build agents.
Feel free to ask me any questions about Kodosumi
--
Whats next?
Kodosumi is just 1 of 3 core pillars of the Sumi ecosystem.
The 2nd big pillar is Masumi: Our Agent-To-Agen Payment Communications Network
And then the 3rd is Sokosumi: Our Agent Marketplace
All 3 pillars are essential to provide our users with the best possible Agentic experience and we're excited to roll those all out over the coming months.
Thanks and have a great day
Kodosumi is a promising open-source runtime environment for running and scaling AI agents effortlessly. Its focus on speed, scalability, and being completely free makes it a go-to option for developers looking to deploy agentic services at scale without the usual complexities.
What stands out is its developer-centric approach—streamlining deployment while remaining fully open-source for flexibility and cost efficiency. For startups and teams working on AI-driven applications, Kodosumi offers a practical and powerful solution.
Though it’s still early days, its clean design and focus on scalable performance show strong potential to become a key tool for AI development teams. Looking forward to its impact!
As builders of the first working prototypes of Kodosumi it was really fun for us to see how excited engineers were about exposing their agents to the public and make use of other agents for their own agentic services, too. Agent to agent communication felt like having a good conversation irl, after sitting for weeks by your own talking to the mirror. Scalable deployment needs to be easy and Kodosumi is helping us get there.
interesting. But you might have to keep evolving quickly as the technology moves. Good one!
Excited to see Kodosumi come to life! How does Kodosumi handle agent scaling for complex, long-running tasks?
Sounds like a mega plan, built an agent recently but was a bit messy & was the first one I'd tried my hand at - Would love the opportunity (once I get better at it) to actually implement some tool like your product :)
Best of luck today Patrick & team, I really look forward to following the product!
kodosumi also works well for any tasks which are intended to run for prolonged periods of time
Kodosumi sounds like a solid foundation for AI agent deployment — fast, scalable, and open-source is exactly what devs need to move quickly without constraints.
