FAQ

Short answers to the questions people ask most about platforms, versions, networking, packages and data.

Short, direct answers to common questions. Each one links to the page with the full detail.

Platforms and requirements#

Which operating systems does VeloPhex run on?#

Studio and the Robot run on Windows x64. Studio needs Windows 10 version 1809 or later, Windows 11, or Windows Server 2019 or later. The Orchestrator server runs as Linux containers on a Linux x86-64 host with Docker Engine 27 or later. See System requirements.

Does the Robot run on Linux or macOS?#

No. The Robot and Studio are Windows x64 applications. Only the Orchestrator server runs in Linux containers.

Which Python version do Python automations use?#

CPython 3.12, and only 3.12. The SDK requires ==3.12.*, and the Robot's Python runtime prepares every environment with its own CPython 3.12 and uv, so an automation behaves the same on your machine and on a Robot. The Python runtime is a separate Robot component, offered only by Setup builds that include it. See Python SDK overview.

Do I need to install .NET on Robot machines?#

No. The Robot service and the workflow executor are self-contained. Studio also carries its own runtime.

Networking#

Which ports do I need to open?#

Robots connect outbound only, over HTTPS (TCP 443) to the Orchestrator. You open no inbound ports on Robot machines. On the Orchestrator host, open TCP 443 to users, Studio and Robots, and TCP 80 if you use automatic Let's Encrypt certificates. The host also needs outbound SMTP for invitation and notification mail. See Network and firewall.

Can Robots connect through our corporate HTTP proxy?#

Yes. By default the Robot uses the machine's WinHTTP proxy (netsh winhttp). You can also set an explicit proxy with a bypass list, and authenticate to it as the machine account or as a named user, with VeloPhex.Robot.Service.exe configure --proxy .... Python restores and the workflow executor get the proxy address but not its credentials, so an automation's own traffic needs a proxy that admits it without them. See Network and firewall.

Can I put the Orchestrator behind my own load balancer?#

Yes. Start the stack without the bundled Caddy service and point your proxy at the api. You then own the proxy settings the Robot connection depends on: WebSocket upgrade forwarding, a long idle timeout, no response buffering on the gateway and hub routes, the original Host header, request bodies as large as the upload limit (100 MB by default) and VELOPHEX_TRUSTED_PROXY_NETWORK. See Install on-premises.

Packages and automations#

Do my packages have to be signed?#

Customer packages do not have to be signed by default. A signature that is present is always verified, and a broken one is always refused. You can require signatures on the Orchestrator and on each Robot once you pin your organization's signing certificate. VeloPhex's own packages must always carry the VeloPhex signature. See Package signing and supply chain.

Which activity packages are supported?#

The core packs are System, UI Automation (desktop and browser), Excel and Orchestrator. The Orchestrator pack carries the queue, asset, job and storage activities and the Action Hub and Data Hub activities. Setup also includes Testing and SAP; these six packs are the supported Core profile. Connector packs are in the activities feed but are not part of the supported profile. See Activities overview and the Activity catalog.

Can a workflow call a Python automation?#

Yes. Publish the Python automation, then invoke it from the workflow. It runs as a step of the same job, on the same Robot, so that Robot needs both runtimes. See Python SDK overview.

Why was my job refused with no_capable_runner?#

No active Robot that serves the workspace reports the runtime the package needs, for example a Python package when only workflow Robots serve it. Assign a machine with that runtime to the workspace. The identifier keeps an older name for compatibility.

What happens if a Robot stops in the middle of a job?#

If the automation had already started, the attempt fails with VXR-OUTCOME-UNKNOWN once its lease expires. It is never retried automatically, because the automation may already have changed other systems: check them, then use Re-run. If the Robot reports its result later, the job is reconciled to the real outcome. Business errors are never retried either; other application errors are retried by the process's retry policy. See Jobs.

Data and security#

Where are my credentials stored, and can VeloPhex staff read them?#

Assets and credentials are stored in the Orchestrator's database, encrypted with AES-256-GCM under a per-tenant data key that is wrapped by a key-encryption key you supply. In an on-premises deployment the key-encryption key is yours alone. A Robot receives a secret only through a one-time credential lease just before it needs it, and never writes it to disk or logs. See Secrets and encryption.

What happens if I lose the key-encryption key?#

Every stored asset and credential becomes unreadable, and no backup can bring them back. Escrow VELOPHEX_KEK and VELOPHEX_TOKEN_PEPPER before you store the first credential. See Backup and restore.

How long is job history kept?#

By default, jobs and queue items for 90 days, execution logs for 30 days, notifications for 90 days and audit events for 365 days. Tenant administrators can change each window within platform limits. See Retention.

Scale and operations#

How many Robots can one Orchestrator handle?#

One cell has been certified with 1,000 connected Robots at 1,000 job starts per minute, with no failures and a median start-to-claim time of 145 ms, on the development topology. Larger fleets are not yet certified. See Sizing.

How do I upgrade without losing data?#

Take a cold backup, pull the new image, apply the release's schema script, then restart. Schema changes are additive, so you roll back by returning to the previous image, never by rolling back the schema. Upgrade the Orchestrator first, then the Robots. See Upgrade.

Where do I find logs?#

See Log locations.

Platforms and requirements

Which operating systems does VeloPhex run on?

Does the Robot run on Linux or macOS?

Which Python version do Python automations use?

Do I need to install .NET on Robot machines?

Networking

Which ports do I need to open?

Can Robots connect through our corporate HTTP proxy?

Can I put the Orchestrator behind my own load balancer?

Packages and automations

Do my packages have to be signed?

Which activity packages are supported?

Can a workflow call a Python automation?

Why was my job refused with no_capable_runner?

What happens if a Robot stops in the middle of a job?

Data and security

Where are my credentials stored, and can VeloPhex staff read them?

What happens if I lose the key-encryption key?

How long is job history kept?

Scale and operations

How many Robots can one Orchestrator handle?

How do I upgrade without losing data?

Where do I find logs?