# How it behaves (what it looks like) A user sends an email to a specific email address, e.g., [email protected]. Kiku was created as a designated coordinator agent that takes the requested task, like “take the attached article and prepare social media posts”, and routes it to one or more internal agents via email. Kiku knows to which agents to route to because “she” maintains a “roster” of the agents in the cluster. This roster describes the capabilities of each agent. Agents act on the part of the task sent to them and return the results to the coordinator who then returns it to the user.
One of the key design aspects is the project uses email threading as state memory. When the agent’s LLM is called, the the thread is passed in as history context.
Top Benefits of the Kikubot framework:
- Users engage AI agents via email which makes it easy for organizational deployment - no end-user installs, just an email address to send tasks too - Per agent user access controls - Per agent containerization - Highly scalable capabilities - agent teams can “nest” other teams. - Cost containment via per agent LLM selection - High visibility into intra-agent “conversations” via standard email tools
Ease of Deployment: We are working on making it easy to deploy and customize. We have a Configurator tool that we believe goes a long way in making setup much easier. We just added a ./demo.sh script that launches a complete demo on your machine included a local email server, webmail client and Kikubot agent. Just run ./demo.sh.
Comparison with OpenClaw: There’s a billion agent frameworks out there. Here is a side by side table comparison with OpenClaw to help understand where Kikubot fits:
Dimension | Kikubot | OpenClaw
-----------------+------------------------+-----------------------
Message bus | Email (IMAP/SMTP) | Chat apps (Telegram,
| | Slack, Discord,
| | WhatsApp, etc.)
-----------------+------------------------+-----------------------
Architecture | Multi-agent by design | Single gateway with a
| - coordinator + | channel/brain/body
| specialized sub-agents | layer separation
-----------------+------------------------+-----------------------
Target scale | Organizational / | Personal productivity
| enterprise (hundreds | / individual assistant
| of agents, |
| departments) |
-----------------+------------------------+-----------------------
Agent topology | Decentralized - each | Local-first gateway
| agent independently | process on your
| deployed anywhere in | machine or VPS
| the org |
-----------------+------------------------+-----------------------
End-user install | None - users just | Requires installing
| email an address | and running the
| | gateway runtime
-----------------+------------------------+-----------------------
Primary language | Go | Node.js