Full control over every agent
Watch progress live. Cancel mid-run. Steer while it works. Set permissions per thread, not per admin ticket.
Chat is where work happens now. FlySpace is a workspace built for working with AI agents: full control over every conversation, documents that grow out of threads, and Claude Code at the core.
Early access rolls out in small batches. No spam, ever.
Deploy the pricing fix and tell me when it's live.
✓ pnpm build (42s)
▸ deploying to production
Live at flyspace.ai/pricing. Want this thread as a changelog entry?
Your AI lives inside someone else's message schema. Interrupting a run, tuning its permissions, even knowing what it's doing right now: all workarounds, all on someone else's roadmap.
The decisions, the reasoning, the answers: trapped in threads that nobody will find again. Your team's knowledge deserves better than an infinite scrollback.
Turning a conversation into a document means copy-pasting it out of the place where the thinking happened. The medium fights the work.
Watch progress live. Cancel mid-run. Steer while it works. Set permissions per thread, not per admin ticket.
Distill a conversation into an article and publish it to your team or the open web, without ever leaving the workspace.
Every thread is backed by a real Claude Code session: your repos, your tools, your skills. Remote, persistent, always reachable.
Run your team's conversations and agents in one place.
A personal command centre for everything you build with AI.
Moderation, answers, shared notes: discussions worth reading.
I ran my consultancy from Slack with a bot I built. It wired Claude Code into every thread, and it changed how fast I could switch contexts and ship.
Then I hit the ceiling. I couldn't cancel runs cleanly, couldn't see where a message had got to, couldn't update my own app past workspace permissions. The medium was the limit.
So I'm building the medium. This page was written, designed and deployed from a FlySpace thread.