Build programs that can read each other
Tau programs can check the other apps, assets and rules in the system before they interact. Write the spec, and the resulting software is correct by construction according to the spec.
Programs that reason about other programs
On Tau, a program can read the meaning of another program, and of itself, before it acts. Everything built on Tau has more powerful abilities.
Every proposal is checked against the DAO’s own rules before it passes, so no vote can drain the treasury or break the constitution.
Tokens and assets hold their own conditions and check the rules of any account or app before they move.
An app sees what another app will do before interacting with it, and only goes ahead when both sets of rules agree.
Contracts are built from their specification and checked against every contract they call, so conflicts are found before deployment, not after an exploit.
Enable rapid development by adding requirements from any number of developers, AI agents and stakeholders and your Software adapts to them collectively, making team growth simple.
Requirements are the new software
Tau Language is a decidable variant of temporal logic, so the correctness of every operation is proven in finite time, without wasting compute.
Each description is an executable specification. Testing no longer hunts for bugs; it confirms the description is what you want.
Build the entire system with full formal verification, synthesized from its specification, reducing error risk and redevelopment costs.
Conflicting requirements are detected automatically, so software stays consistent as rules and regulations change.
Upgrades without forks or downtime
The whole system is formal, so every upgrade is guaranteed to meet its specification.
When users agree on an update, it goes into the blockchain and everyone updates from one block to the next. When they don’t, logical agree/disagree detection shows exactly where they differ, so a real hard fork becomes rare.
The system changes only according to its own rules for change, agreed by consensus. Users collectively define what consensus means to the system.
