RenderFlow 2 is out

Today we’re releasing RenderFlow 2, our render farm manager for visualization, animation and VFX. It’s built for studios that render on their own machines, and shaped by a year of conversations with them: at FMX, at SIGGRAPH and in countless meetings about what an ideal render farm manager and network rendering should look like.

RenderFlow 2 runs on your own network and your own server, on Windows, Linux and macOS, from one machine to more than a thousand. Here’s what’s new.

In-app submission

Send a scene to the farm from inside the app you work in. 17 applications have a submitter, from 3ds Max, Maya, Cinema 4D and Blender to Houdini, Nuke, After Effects and Unreal Engine. You can also pick a scene file and submit it without opening it, and get the same job as from inside the app.

Multi-step jobs

A job can do a lot more than render. Export a scene, run your post-production, encode the video or run a script, all in one submission. Each step starts when the one before it finishes, with the files it made.

Batch submission

One submission can create multiple jobs. Every Cinema 4D take, Maya render layer, After Effects render queue item or Unreal Engine Movie Render Queue job in a scene becomes its own job on the farm.

Live render preview

A thumbnail for every finished frame and a live preview of the one still rendering, so a bad frame shows up before the whole sequence is done.

Artist-friendly logs

When a frame fails with a known error, the log explains it in plain language and says what to try next. RenderFlow knows the errors of the applications it supports, and your TDs can add the studio’s own errors and fixes to the catalog, so artists can sort things out themselves.

Smart job matchmaking

Jobs go to the machines that have the software and memory to render them, fastest first. A machine that can’t take a job says why, and a job nobody can render waits and tells you why. Benchmarks score every machine, so the work goes to the fastest ones first.

Overnight rendering

When an artist leaves for the day, their workstation joins the farm, and it goes back to them when they return. Machines with nothing to render switch off, and wake up when work arrives.

Hybrid cloud in your own account

Need more power? Connect your own Azure or AWS account. Your own machines still render first. Cloud machines start only when those are all busy, and leave when the work is done. There’s no Pulze cloud service in between: the machines and the bill stay in your account.

Notifications

Get a message when your job finishes or runs into an error: by email, on the desktop, or in Slack, Teams or Discord.

Farm view

Every machine and every job on one canvas, with live CPU, GPU, RAM and VRAM readings for each machine.

Statistics

See who renders what, and when: an activity heatmap, utilization, and render time per machine and per person. The same data comes through the API for your own dashboards.

Scheduler

Plan the farm’s week. Suspend work, move machines between pools, shut them down, wake them on time or run a script, all on one weekly calendar.

Inventory

Every application and plugin version on every machine, compared against a reference machine, so a wrong version stands out before it breaks a render.

MCP server

RenderFlow has a built-in MCP server, so Claude or any AI assistant that supports MCP can connect to your farm. Ask why last night’s job failed or which machines are free right now, or tell it to render the failed frames again. An administrator decides whether it’s switched on.

Ready for IT

Silent and headless installs, TLS with your own certificate, permissions per group across 48 features, API keys with their own permissions, and an audit log of who did what. SSO, SCIM and offline licensing (Enterprise only).

For developers

Everything in the interface is also available through a REST API, Python and TypeScript SDKs, the rfcli command line and webhooks. Write a Python plugin and it becomes a new job type for the whole studio.

Coming from RenderFlow 1?

RenderFlow 2 starts from an empty database: jobs, settings and history from RenderFlow 1 don’t carry over, and scripts that use the RenderFlow 1 API need updating. The migration guide has the steps.

Try RenderFlow 2

Try it out today on up to 10 machines for 30 days, and see how RenderFlow handles your own scenes. Every change, including the fixes, is in the changelog.