Trigora Launches Durable Execution Platform Using TCC to Avoid History Replay

Trigora has released a durable execution platform designed for long-lived AI agents and programs. Unlike traditional workflow systems that reconstruct state by replaying history, Trigora employs Transparent Continuation Checkpointing (TCC). This method commits the program position and live durable state at boundaries, allowing recovery to resume from a committed continuation rather than re-executing completed steps.

Core Technology and Use Cases

The TCC engine is implemented in Rust and supports native and WebAssembly embedding. It is specifically targeted at workloads that:

• Run long-lived AI agents calling models, tools, and APIs.
• Wait for human input or external events.
• Sleep for minutes, hours, or days.
• Coordinate durable child executions and dynamic branching.
• Require survival of worker or process failures.

Language Support and Deployment

Developers can write application logic in TypeScript, Python (3.10–3.14), or Rust. The ecosystem includes SDKs, clients, and a CLI for each language. Key features include:

• Local Development: Initialize projects with trigora init and run locally with trigora dev.
• Production Deployment: Deploy via trigora deploy or use Trigora Cloud, a managed runtime offering events, timers, cron/webhook triggers, and observability.
• Licensing: The Trigora SDKs, CLI, and public contracts are MIT licensed. The underlying TCC Engine is source-available under Business Source License 1.1, converting to Apache 2.0 under specific terms.