cronflow

Cronflow executes workflow steps through a Rust core compiled via napi-rs that bridges to Node.js/Bun runtimes, eliminating JavaScript interpretation overhead for performance-critical operations. The engine manages job dispatch, worker pool scheduling, and state transitions at the native binary level, achieving sub-millisecond execution latency by avoiding serialization costs between TypeScript definitions and execution. This hybrid architecture allows TypeScript to define workflows declaratively while Rust handles the actual execution, persistence, and scheduling logic.

Workflows are defined as TypeScript code using a fluent builder API (e.g., `workflow.step().if().parallel().while()`) rather than JSON/YAML configuration, enabling version control, unit testing, and IDE autocomplete. The SDK provides type-safe step definitions with Zod schema validation for payloads, allowing developers to catch errors at compile-time rather than runtime. This approach treats workflows as first-class code artifacts, not configuration files, integrating with standard software engineering practices.

Cronflow manages concurrent step execution through a Rust-based worker pool that dispatches steps to available workers, with configurable pool size and parallelism limits. The worker pool is implemented in the Rust core, avoiding JavaScript event-loop contention and enabling true parallelism. Steps are queued and executed as workers become available, with the engine managing synchronization and result aggregation.

Cronflow supports triggering workflows via HTTP webhooks (with built-in or external webhook servers), cron-based schedules (via Rust scheduler), and custom application events. The trigger system is implemented at both the Rust layer (for performance-critical scheduling) and TypeScript SDK layer (for webhook registration and event binding). Webhooks integrate with Express, Fastify, Koa, and NestJS frameworks, allowing workflows to be triggered from existing web applications without additional infrastructure.

Workflows support imperative control flow constructs including conditional branching (if/else), parallel step execution, and while loops, all defined in TypeScript and executed by the Rust core. Parallel steps are dispatched to the worker pool simultaneously, with the engine managing synchronization and result aggregation. This allows complex business logic to be expressed directly in workflow definitions without external orchestration logic.

Workflows can be paused at any step to await manual approval, with the engine generating cryptographic tokens that authorize resumption. The paused state is persisted in the Rust core, allowing workflows to survive application restarts. Approval tokens are time-limited and can be validated before resuming execution, enabling secure human-in-the-loop automation for sensitive operations like deployments or financial transactions.

Cronflow provides pre-built webhook server integrations for Express, Fastify, Koa, and NestJS, allowing workflows to be triggered from HTTP requests without running a separate webhook service. The SDK registers webhook routes that validate incoming payloads against Zod schemas and dispatch them to the Rust core for execution. This enables workflows to be embedded directly into existing web applications.

Cronflow persists workflow state (including paused workflows, execution history, and step results) in the Rust core using a binary format optimized for performance. State is automatically managed across workflow executions, allowing workflows to resume from checkpoints and maintain context across multiple invocations. The persistence layer is abstracted from the TypeScript SDK, requiring no external database configuration.

Cronflow provides pre-compiled native binaries for Windows, macOS, and Linux (x86_64 and ARM64), eliminating the need for users to compile the Rust core locally. The build pipeline uses napi-rs to generate platform-specific binaries that are packaged with the npm package, allowing installation via `npm install` or `bun install` without requiring a Rust toolchain. This approach makes Cronflow accessible to developers who don't have Rust experience.

Cronflow uses Zod schemas to define and validate workflow payloads and step parameters at runtime, providing type safety and runtime validation without requiring TypeScript compilation. Schemas are defined in TypeScript and automatically generate type definitions, enabling IDE autocomplete and compile-time type checking. Validation errors are caught before workflows execute, preventing invalid data from propagating through steps.

Cronflow exposes performance metrics including step execution latency, workflow throughput, and worker pool utilization, collected by the Rust core and exposed via the TypeScript SDK. The engine tracks sub-millisecond execution times and provides benchmarking utilities for measuring workflow performance. Metrics are available in-process without external monitoring infrastructure.