What's New

Hosted Functions for JavaScript, Python, and Golang

Build custom workflow logic directly in Unmeshed using JavaScript , Python, and Golang. Run scripts in isolated containers that scale for high-volume execution while keeping orchestration and operations in one place.

Get Started

Script directly in workflow steps

JavaScript, Python, and Golang steps accept inline scripts and return structured step results. Each step has access to upstream step data and process context, making transformations and decision logic easy to implement.

  • Build custom business logic quickly with JavaScript, Python, or Golang in the same orchestration platform.
  • Reuse upstream workflow data and process context without writing separate integration glue.
  • Keep execution behavior consistent across languages with a unified step contract and observability model.
  • Ship faster by running language-specific functions under one reliable, production-ready orchestration runtime.

JavaScript Function

javascript.step.js

(steps, context, helpers) => {
  const locationId = context.input.locationId;
  const data = steps.user_fetcher_ref.output.response;
  const filteredData = data.filter(
    (it) => it.locationId === locationId
  );

  return {
    id: locationId,
    locations: filteredData,
  };
};

Python Function

python.step.py

import os

def main(steps, context):
    file_path = "/app/files/10mb.pdf"

    if not os.path.exists(file_path):
        return {
            "processId": context.get("id"),
            "currentStepId": steps["__self"]["id"],
            "isSuccessful": False
        }

    with open(file_path, "rb") as f:
        binary_data = f.read()

    return {
        "isSuccessful": True,
        "sizeBytes": len(binary_data)
    }

Isolated containers, built for scale

JavaScript functions, Python scripts, and Golang steps run in isolated containers managed by Unmeshed. This model supports high-volume executions while preserving predictable runtime boundaries and operational reliability.

  • Containerized runtime isolation per execution type.
  • Autoscaling model for sustained high-throughput workloads.
  • Strong fit for API-heavy, event-driven, and data-processing workflows.

Python file workflows and runtime extensibility

Python steps can access files managed in Unmeshed directly from /app/files, making them ideal for document, report, and large-input workflows. For advanced workloads, Python containers can be expanded to install additional libraries including numpy, pandas, and third-party packages such as LangChain.

Native file access

Read platform-managed files from /app/files in Python script tasks by default.

Extended libraries

Expand Python containers to include ecosystem libraries for analytics, AI tooling, and data science.

One orchestration plane

Run mixed-language business functions in one orchestrated, observable execution platform.

Teams shipping powerful workflows chooseUnmeshed

Ready to run JavaScript and Python business logic in isolated, autoscaling containers? Build and operate hosted functions in one orchestration platform.

Get Started