Thanks, Computer is a distributed runtime for coordinating business logic across systems, services, humans, and time.
Where an operating system coordinates processes on a single machine, TxCo coordinates operations across an event flow:
● event in
│
┌──┴──┐
│flow │
└──┬──┘
┌─────┼─────┐
▼ ▼ ▼
┌───┐ ┌───┐ ┌───┐
│op │ │op │ │op │ parallel execution
└─┬─┘ └─┬─┘ └─┬─┘
└─────┼─────┘
merge
│
▼
● output
An incoming event (HTTP, email, cron, queue, webhook, etc.) enters an operation stack (OpStack).
At each step, TxCo evaluates resonators — small rules written in txcl that determine which operations should run. Matching operations execute in parallel, their outputs deep-merge into a shared event document, and the flow continues to the next step.
TxCo separates three concerns:
- Flow — when and under what conditions operations run
- Execution — the actual code or external service being invoked
- Merge — combining results back into the event flow
This separation makes it possible to coordinate workflows that span:
- multiple services
- multiple organizations
- humans and AI systems
- asynchronous and long-running processes
Operations stay small and composable, while the OpStack defines how they cooperate as a larger system — similar to how departments coordinate work inside an organization.
TxCo draws inspiration from early distributed AI systems and event-driven architectures, particularly:
- Blackboard Architecture
- Oliver Selfridge's "Pandemonium" model of independent cooperating agents
These systems explored how complex behavior can emerge from many small, specialized processes operating over shared state.