Kodosumi

Kodosumi

Runtime environment to execute agentic services at scale

5.0•2 reviews•

221 followers

Visit website
Deploy and scale your AI agents with Kodosumi, the open-source runtime built for developers. Fast, scalable, and fully free to use.
This is the 2nd launch from Kodosumi. View more

Kodosumi 1.0.0

Run AI Agents at Scale, Reliable and Fast
The distributed runtime environment that manages and executes agentic services at enterprise scale. Fast, scalable, and fully free to use.
Kodosumi 1.0.0 gallery image
Kodosumi 1.0.0 gallery image
Kodosumi 1.0.0 gallery image
Kodosumi 1.0.0 gallery image
Free
Launch tags:Open Source•Developer Tools•Artificial Intelligence
Launch Team / Built With
AppSignal
AppSignal
Built for dev teams, not Fortune 500s.
Promoted

What do you think? …

Jami Safari
Sokosumi
Hunter
📌
Hi Hunters, I’m Jami, part of the team behind the Sumi ecosystem. We built Kodosumi to make deploying and scaling AI agents simple. Many devs and domain experts who are not infra engineers hit the same wall we did: long agent runs, APIs, scaling. Kodosumi handles that so you can focus on the agent itself. With v1.0.0, we’re shipping some big updates: - File upload and download - Synced documentation across all public features - Pytests on the most important functions - Basic health checks - More examples in kodosumi-examples - Early HITL (human-in-the-loop), still experimental This release marks the shift from prototype to a stable foundation you can rely on for production. Kodosumi is one of three pillars in the Sumi ecosystem: Masumi (agent to agent payments and communication) Sokosumi (agent marketplace) Together, they form the stack we believe will power the agentic economy. Would love to hear your thoughts and see what you build with v1.0.0.
Sneh Shah

Congrats on the launch, team! 🚀 As developers tackle ever-more complex AI stacks, Kodosumi stands out for simplifying agent deployment and scaling. Can you share how your distributed runtime optimizes routing for different agentic services (LLMs, workflows, providers, etc.), and what efficiency gains you’ve seen for latency, cost, or throughput at scale?

Also curious about your foundation for stateful, memory-aware agents—how does Kodosumi’s approach to context management help agents maintain cross-session memory or support more advanced workflows?

Michael Rau

@sneh_shah kodosumi is framework agnostic. So how you route your agentic service depends on the framework you chose. What kodosumi brings to the table is a start+stream+final-response control. scaling builds on top of ray as a means to distributed execution. Similar to cross-session memory store. This (should) ship with your agent framework and use in-memory or persistent stores as for example ADK is doing it with session services

Phil Isenmann

Kodosumi is an amazing product!

Katsche

Heyho frens!

This is a big upgrade. We experience that daily: drafting an agent is one thing. But putting it really to work and integrate it into a company is totally different issue. Ray clusters are magic but with kodosumi these become available for everyone super fast and easy. The upgrade were desparately expected and bring kodo to the lastest state-of-the-a(i)rt. Thx so muchm, team.

Just image what v 4.0 will look like!?

K