AI-generated editorial illustration; not a documentary photograph.
LittleHorse has released Saddle Command Center 1.3, an update to its workflow orchestration platform that adds no-code agent deployment, prebuilt task workers, a JavaScript SDK and a free serverless trial. The company says the platform coordinates AI agents with conventional software across application, API and event-driven workflows.
More control around agents
The release centers on a decision-worker pattern. An agent can receive read-only access to tools and handle a judgment-heavy step, while deterministic workflow code validates the result and performs the resulting action. LittleHorse says the arrangement is intended to limit the agent's authority and make failures easier to inspect.
Users can configure an agent by selecting Model Context Protocol servers, setting a system prompt and defining input and output structures. The platform deploys that configuration as a task worker, records tool calls and exposes a stack trace for debugging. Workflow rules handle business errors, while policies can retry technical failures such as network problems.
Broader workflow runtime
Saddle Command Center 1.3 also includes an Apache Kafka-compatible streaming engine, webhook and data-source connectors, and a runtime for prebuilt agents and task workers. Existing code can be connected as workflow components through APIs, event queues or custom tasks. Java, Python, Go and C# were already supported; JavaScript is new in this version.
LittleHorse describes its approach as “Business-as-Code”: business processes are represented as version-controlled, testable graphs that technical and business users can inspect and modify. The system is designed for processes that span applications and may wait for external responses, including a merchant-onboarding workflow that the company said could take up to three days because of identity-service dependencies.
Availability and business model
The free serverless option is intended to let developers test the platform without provisioning infrastructure, and version 1.3 is available now. LittleHorse uses an open-core model, with an open-source orchestration kernel supporting its commercial platform.
The privately held company said its software has been in production for more than 18 months, has paying customers in multiple industries and is expected to book more than $1 million in revenue this year. Those figures are company statements, not independently verified results. Enterprise pricing is based on processing capacity, including Kafka throughput and workflow tasks per second.
SIGNAL TITANS
Leave a Reply