prose

You can orchestrate multi-agent workflows by writing declarative OpenProse programs that define agent configurations, sequence tasks, and manage state. This approach replaces simple prompts with structured code for robust, repeatable AI processes.

State persistence for long-running agent tasks is managed by storing execution state across file-system, SQLite, or PostgreSQL backends. This ensures scalable and inspectable execution for complex operations.

Yes, you can sequence and parallelize LLM agent tasks by writing .prose programs. The declarative language lets you define reusable agent configurations with specific models and prompts to automate complex workflows.

OpenProse is a declarative programming language for AI agent orchestration that enables the definition and execution of multi-agent workflows. It facilitates complex task automation through structured code rather than simple prompts.

Yes, the tool supports SQLite and PostgreSQL as state persistence backends alongside file-system storage. These options allow you to persist execution state for long-running or complex multi-agent operations.

Automate code review by writing a .prose program where one agent identifies bugs, another checks style compliance, and a third synthesizes feedback into a single report. This sequences reusable agent configurations into an automated workflow.