Kalceo is one of those rare projects that makes the agent hype feel earned. It's a B2B SaaS being built by an autonomous AI agent fleet as part of Ekioo's R&D program, and the whole thing is documented in public — not as a demo, but as a production operation. The target market is the construction industry, which still runs on legacy tooling and manual workflows that are begging for automation. Kalceo doesn't exist in isolation. It sits inside Ekioo's broader agent-fleet experiment alongside Bloomii, described as a constructive-journalism media project, and KittyClaw, the kanban orchestrator that coordinates all of the agents running under the hood. That trio is what makes this worth reading: you're not getting one toy pipeline, you're getting a fleet of specialized agents each pulling a different workload. The most interesting part for builders isn't the novelty — agent fleets are everywhere now. It's the documentation angle. Ekioo is publishing how these systems actually operate in production, which is still rare and genuinely valuable. Most teams show you the happy-path demo; this is closer to watching the real machinery run.

The Fleet in Production

KittyClaw deserves special attention here because it's the orchestrator — the piece that decides which agent does what, when, and how work flows between them. If Kalceo is the product and Bloomii is the media experiment, KittyClaw is the nervous system holding the whole operation together. That architecture matters for anyone trying to build multi-agent systems that don't collapse under their own complexity. What's missing from the write-up so far are hard numbers: how many agents, what models they're running on, throughput metrics, cost per task. The post frames this as ongoing R&D rather than a shipped product with benchmarks, and it would be nice to see concrete performance data in future installments. For now, treat Kalceo as evidence that the approach works at all — not as a template for production scale.

Key Takeaways

  • Kalceo is a B2B SaaS targeting construction, built by Ekioo's autonomous agent fleet alongside Bloomii (media) and KittyClaw (kanban orchestration).
  • The entire operation runs in production and is publicly documented — a rarity in the AI-agent space.
  • KittyClaw acts as the orchestrator coordinating all agents, making it the architectural backbone of the fleet.

The Bottom Line

This is what real agent-fleet development looks like: boring, operational, and shipping actual software instead of demos. Ekioo is doing the unglamorous work that most AI teams avoid — documenting production behavior in public. If you're building multi-agent systems, Kalceo is worth a close read.