Product

What you can build.

Long-lived programs that suspend, survive worker loss, and continue from a committed continuation without replaying completed execution history.

Trigora Cloud

Run TCC without running the infrastructure.

Trigora Cloud is the managed production platform for deploying Programs, running Executions, and handling durable recovery.

Deploy

Ship Program versions without operating the execution infrastructure yourself.

Observe

Inspect running, waiting, sleeping, failed, and completed Executions.

Integrate

Start, signal, and resume Executions through APIs, events, timers, child activations, and external systems.

Start building →

Programming model

Write ordinary programs.

Most application logic runs as normal code. Durable operations are where execution crosses a boundary.

  • Control flow Conditionals, loops, exceptions
  • Data & computation Values, arithmetic, objects, arrays, helpers
  • Structured concurrency Promise.all / gather / join, Promise.race / race
  • Durable operations Effects, waits, timers, child executions

Concurrent branches remain durable across suspension and recovery. TypeScript Promise.all, Python gather, and Rust join preserve input order. Promise.race and race persist the first durably accepted winner.

  1. ordinary control flow

    values · helpers · if / else · loops · try / catch

  2. durable boundary

    effect · wait · sleep · invoke

  3. committed continuation

    live state at the boundary

  4. resume after interruption

    load checkpoint — do not replay the prefix

Durable boundaries

Durability is part of execution.

Effects, waits, child executions, and other durable operations create points where execution can safely survive interruption.

Effects

External work whose committed outcome can be reused during recovery.

Waits

Execution can suspend without keeping a process alive.

Children

Long-running work can compose into durable execution trees.

TCC Engine

Transparent Continuation Checkpointing

The portable engine underneath Trigora Cloud. Checkpoint at durable boundaries; recover by loading the continuation — not by replaying history.

Recovery

The process is disposable. The execution isn't.

Once continuation state is committed, another process can restore that execution and continue from the saved boundary.

Recovery from a committed continuation checkpoint does not require prefix replay.