Blog
Stories from inside the agent.
How AI-native workspaces work, how agent memory lasts, how privacy survives when agents act, and how Kylon compares.
Fold, don't forget
An agent is a long-term member of your team. The conversation never ends, but the context window never grows. Here's how we solved it: by folding history into a tree that can always be unfolded again.
A User Message Is More Than a User Message
When a user types a message in an AI-native workspace, the system sees far more than text. We unpack Kylon's activation architecture and show how a single message becomes a structured execution contract.
Why Privacy Is the Hardest Problem in AI Workspaces, and How Kylon Solves It
When AI agents operate with real agency inside a workspace, privacy becomes an architecture problem. This post walks through how Kylon keeps agents confined to authorized data, even under adversarial conditions.
Kylon vs. Claude Tag: Same idea, very different architecture
Anthropic just put Claude inside Slack. We've been building AI teammates into our own workspace from day one. Here's what that difference actually means.
Why we built Kylon
Most AI tools are a tab you open alone. We wanted agents that join the room — that read the thread, follow the work, and ship.