VeloPhex Orchestrator

The control plane for your automation estate: tenants and workspaces, jobs, queues with SLAs, triggers, encrypted assets and credentials, a package feed per tenant, Robot fleets and an append-only audit log. Run it on-premises, in the VeloPhex cloud or in a dedicated cloud managed by VeloPhex.

Tenants and workspaces

Each tenant has its own encryption key and package feed; workspaces separate teams inside it.

Runs where you need it

On-premises with Docker Compose, in the VeloPhex cloud or in a dedicated cloud.

Part of one platform

Governs the Studio and Python packages you publish and the Robots that run them, through one REST API.

No shared view of automation

Every Robot, queue and job in one control plane.

Missed deadlines

Queues with priority, deadlines, SLA and risk-SLA.

Audit requests

Append-only audit events for identity, roles, assets, packages, jobs, queues and machines.

Deployment constraints

Keep Orchestrator in your own data center or cloud account when policy requires it.

Automation CoE and platform owners

Run the whole estate of jobs, queues and Robots from one place.

IT operations

Watch fleet health, queue SLAs and failures, with traces over OpenTelemetry.

Security and audit teams

Review who changed what, with role-based access and an append-only audit log.

Request a demo Check automation fit

Queues & SLAs

Work queues with priority, deadlines, SLA and risk-SLA, and retries for application errors.

Schedules & triggers

Cron triggers with time zones and queue triggers, run by a built-in scheduler.

Robot fleet presence

Heartbeats every 15 seconds, fleet summaries and live status in the browser.

Role-based access

Built-in roles and groups, assigned at tenant or workspace scope.

Encrypted assets and credentials

AES-256-GCM envelope encryption with a data key per tenant.

Audit trail

Package feed per tenant

A NuGet v3 feed that works with dotnet and CI using a personal access token.

REST API and webhooks

A documented REST API (OpenAPI) and signed webhooks with retries.

OpenTelemetry

Traces and metrics over OTLP, with correlation IDs across requests.

Managed AI agents

Define an agent once, publish it as an immutable version and run it as a governed job, with per-run limits on model calls, tokens and time.

Approvals for agent tool calls

Mark a tool as needing approval. Orchestrator pauses the run until a person approves that one call, and records who decided.

AI gateway: bring your own model

Route agent model calls through Orchestrator to OpenAI, Azure OpenAI, Anthropic or an OpenAI-compatible endpoint such as Ollama or vLLM, with usage and budget tracking.

MCP and automation tools for agents

Give agents published automations, MCP servers, HTTP calls, vector search and human tasks as tools, each inside an allow-list.

Agent evaluations

Score agent versions against evaluation datasets, including LLM-judged checks, and block a publish that falls below the threshold.

Agent traces

Follow every model call, tool call and approval in a run as a trace, live or after the fact.

Agent memory and skills

Memory that carries context across runs, and reusable skills an agent can load when it needs them.

More model providers

Amazon Bedrock and Google Gemini behind the AI gateway.

Audit streaming and export

Stream audit events to Splunk, Datadog, syslog over TLS or an HTTPS webhook, and export them as CSV.

Teams and Slack alerts

Job, queue and agent alerts delivered to Microsoft Teams and Slack channels.

Orchestrator high-availability mode

Multiple gateway replicas and a leader-elected worker, validated for failover.

Enterprise Automation

Automate end-to-end processes across ERP, CRM and legacy systems with governance built in.

Hybrid Automation

Run Orchestrator on-premises or in a dedicated cloud, with Robots next to the systems they automate.

Python Automation

Give developers a real Python framework for automation, with the operations layer included.

Automation Modernization

Move existing bots to a modern platform one process at a time, without a big-bang rewrite.

Can I self-host VeloPhex?

Yes. You can run Orchestrator on-premises with Docker Compose. VeloPhex can also run it for you in the VeloPhex cloud or in a dedicated cloud. Robots always run on your own Windows machines.

Do I need to know Python?

No. Studio is a visual designer, with C# expressions where you need them. The Python SDK is there when you want to write code.

Which operating systems do robots support?

Robots and Studio run on Windows x64. Studio needs Windows 10 version 1809 or later, or Windows Server 2019 or later. Orchestrator runs on Linux x86-64 in containers.

VeloPhex Studio

The Windows IDE for building automations: a visual designer for sequences, flowcharts and state machines, C# expressions with IntelliSense, a live debugger, UI element capture, Git and publishing to Orchestrator.

VeloPhex Robot

The Windows agent that runs workflow and Python automations, next to a person or in the background. It connects to Orchestrator outbound only, proves its identity with its own key pair and verifies every package before it runs.

VeloPhex Python SDK

Write automations as typed Python functions. The SDK turns type hints into input and output schemas, gives your code assets, secrets, credentials and queues at run time, and packs it into the same package format as Studio workflows.

VeloPhex Launchpad

The automation pipeline in Orchestrator. People submit ideas, reviewers assess and score them, approvers decide, delivery teams build them, and Launchpad measures the benefit from the automation's own runs.

VeloPhex QA Hub

Quality management inside Orchestrator: releases, requirements, versioned test cases, test sets, executions on your Robots and defects, with coverage and reports in one project.

VeloPhex Action Hub

When an automation reaches a decision it should not make alone, it raises a typed task. Someone claims it, fills in the form and picks an action, and the automation gets the decision as data and carries on.

VeloPhex Flow Studio

Draw BPMN flows that combine robot tasks, people's tasks, timers and messages. A flow waits between steps for minutes or days without holding a robot, and every published version is pinned for the instances that run it.

VeloPhex Flow Monitor

See how many flow instances are running, which stopped at an incident and how each step performs against its target, then resolve incidents, retry in bulk or migrate instances to a newer version.

VeloPhex Agents

Build an agent from instructions, a model and the tools it may use: your processes and your vector stores. It runs as a job on your Robots, records every step and stops for a person's approval before a process tool call.

VeloPhex Data Hub

Keep business records such as invoices, vendors or claims in typed entities in a workspace. Workflows, Python automations, people and other systems all read and write them through one API.