Skip to main content

Upload and Download Files on Super Renders Farm


Upload and Download Files on Super Renders Farm
Upload and Download Files on Super Renders Farm

Introduction

Moving project files to and from a render farm is the part of the workflow that most often goes wrong. A render fails not because the engine misbehaves, but because a texture path didn't resolve on the remote node, a project went up as a ZIP the farm never unpacks, or a slow home connection stalled mid-upload after twenty minutes.

Super Renders Farm supports three ways to move project files (the web interface, the SuperRenders Client App, and S3 access with Cyberduck or the AWS CLI) and three ways to retrieve rendered output. None is "the right one" in the abstract; each has a sweet spot defined by file size, network speed, and how often you'll re-submit the same project.

This guide gives you a decision tree to pick the method that matches your job, step-by-step instructions for each, and a reference for the upload errors we see most often in support.

Decision Tree: Which Upload Method Should You Use?

The three methods are not interchangeable. Use this short decision tree the first time you submit a project, then switch as your workflow stabilizes.

Use the Web UI if:

  • The total project size is under ~2 GB
  • You submit jobs occasionally (a few times a month) rather than continuously
  • You want zero installation โ€” just a browser

Use the Client App if:

  • You submit jobs at least weekly and want fewer manual steps
  • Your project is larger than ~2 GB
  • You want the upload to retry automatically on connection drops
  • You want completed frames to auto-download to a local folder as they finish

Use S3 access (Cyberduck or the AWS CLI) if:

  • You already work with an S3 client and prefer it to a browser or the Client App
  • You want to script uploads to, or downloads from, your SRF Space; the AWS CLI runs from a terminal or a scheduled script
  • You are comfortable handling an access key (treat it like a password)

A small caveat applies to every path: the farm does not extract archives (.zip, .rar, .7z, .tar, .tar.gz). The farm needs an uncompressed project folder (the scene file plus all referenced assets in the correct directory structure) so the render manager can resolve asset paths. We expand on this in the Common Upload Errors section below.

Method 1: Web UI Upload

The web upload is the simplest path and the one we recommend for first-time submissions and small projects.

Step-by-step:

  1. Sign in to your account at superrendersfarm.com and open the render dashboard.
  2. Click New Job (or Submit Project, depending on which engine landing you arrived from).
  3. Drag the project folder onto the upload area, or click Choose Folder and select it.
  4. Wait for the upload progress bar to complete. The browser shows individual file progress for projects with many assets.
  5. Once the upload finishes, configure render settings (frame range, output format, priority) and submit the job.

Practical limits. Browsers throttle individual upload sessions, so the web path becomes slow above ~2 GB and unreliable above ~5 GB on residential connections. If your browser tab is closed mid-upload, the transfer aborts โ€” there is no resume on the browser side. For projects in that range, switch to the Client App.

What gets uploaded. Drag-and-drop preserves your folder structure exactly. If your scene references textures in \textures\ and \proxies\, those subfolders need to be inside the dragged top-level project folder, with the same names. Missing references on the farm almost always trace back to one of these subfolders being outside the upload root.

Method 2: SuperRenders Client App

The Client App is the path most production studios eventually settle on. It handles resume-on-disconnect, integrity checks, and auto-download of completed frames โ€” none of which the browser path supports.

Step-by-step:

  1. Download the Client App from the Download page and install it on your workstation.
  2. Sign in with your Super Renders Farm account credentials.
  3. Click New Project and select your project folder. The app scans for the scene file and referenced assets.
  4. Review the file list โ€” the app shows missing references before upload starts, so you can fix them locally rather than after a failed render.
  5. Click Upload. Progress is shown per file, and the upload resumes automatically if your connection drops.
  6. After upload completes, configure render settings inside the app and submit.

Auto-download. The Client App also handles the return trip. Once you enable a local output folder, completed frames stream down to your workstation as each frame finishes โ€” you don't have to wait for the whole job to complete before retrieving anything.

Plugin v1 vs v2 note. Studios using older versions of the Client App should check the tools and Client App reference for current install paths, plugin version compatibility, and the migration plan when Spaces (our next-generation transfer layer) supersedes the Client App. That doc covers per-DCC plugin integration, end-to-end submission, and troubleshooting steps in depth โ€” this page only covers the upload-and-download surface.

Method 3: S3 Access (Cyberduck or the AWS CLI)

S3 access lets you move files between your workstation and your SRF Space with a standard S3 client, without the browser or the Client App. Files you upload this way appear in your SRF Space and can be used for a job like any other upload. S3 access moves files only: you still submit the job from the web dashboard, the Client App, or a DCC plugin.

Step-by-step:

  1. In your account, open Cloud Direct Connect and generate your access key (the portal calls this your FTP account). Copy the Access Key ID, the Secret Access Key and the Remote Directory.
  2. Cyberduck: choose Open Connection, select the protocol Amazon S3, and enter the Access Key ID and Secret Access Key. Expand More Options and enter your Remote Directory in Path, then connect.
  3. AWS CLI: run aws configure, enter the Access Key ID and Secret Access Key, set the region to ap-southeast-1, and leave the output format blank. Then copy the unpacked project folder into your Remote Directory, for example aws s3 cp ./my-project s3://<Remote Directory>/my-project --recursive.
  4. Submit the job from the web dashboard, the Client App, or a DCC plugin, picking the files you uploaded.

Getting output back over S3. Finished frames are written to the SuperRendersOutput/<job ID>/ folder inside your SRF Space, so you can pull them with the same client, for example aws s3 sync s3://<Remote Directory>/SuperRendersOutput/<job ID>/ ./frames/.

Good to know. Your account has one access key, and it does not expire. Treat it like a password: keep it out of shared scripts and repositories. If you need it changed, contact support. Temporary and system files (*.tmp, *.temp, *.partial, *.swap, *.lnk, *.ini, *.db) and any rendertemp/ folder are not passed to the render nodes, so keep scene data out of files with those names.

Use Cyberduck with the Amazon S3 protocol or the AWS CLI; FileZilla and FTP/SFTP clients will not connect.

Comparison: Methods at a Glance

MethodBest forPractical size rangeResumes on disconnect?Auto-download output?
Web UIOccasional, small projectsUnder ~2 GBNoNo (manual download)
Client AppStudios with regular submissions and large projects2 GB and upYesYes
S3 access (Cyberduck / AWS CLI)Teams already using an S3 client, or scripted transfersNo limit set by usHandled by the clientYes, if you script the download from SuperRendersOutput/

Match the method to the project shape. A 200 MB still is fine in the browser; a 60 GB animation with cached fluids should go through the Client App, or through S3 access if your pipeline already scripts its transfers.

Common Upload Errors

These are the errors we see most often in support, and what to check first.

"Uploaded a .zip, .rar or .7z and the job can't find the scene."

You can upload .zip, .rar or .7z files, but the farm does not extract them, so nothing inside an archive gets rendered (the web uploader warns that the file won't be unpacked automatically). The same goes for .tar and .tar.gz. Extract your scene folder and upload the files unpacked. If you're working from a colleague's ZIP delivery, unpack it locally first, then upload the resulting folder.

"Missing texture: D:\projects\..."

Your scene references an absolute path on your local machine that doesn't exist on the farm's worker nodes. This is the single most common cause of failed jobs, and it surfaces whether you uploaded via the Web UI or the Client App. Fix it locally by using relative paths (most DCCs have a "make paths relative" or "consolidate assets" command โ€” File โ†’ Reference Editor โ†’ Make Relative in Maya, File โ†’ Archive in 3ds Max, File โ†’ Pack External Data in Blender). The Client App pre-scans for absolute paths and flags them before upload starts, which saves a round-trip.

"Upload timed out."

Common on residential connections, especially over Wi-Fi with files above 1 GB. The browser path doesn't resume โ€” switch to the Client App, which recovers from transient connection drops. Also check your local upload speed (most ISPs cap residential upload at 10โ€“50 Mbps); a 20 GB project on a 10 Mbps line takes roughly five hours of continuous transfer.

"Invalid project structure."

The folder you uploaded doesn't contain a scene file the farm recognizes (.max, .ma/.mb, .c4d, .blend, .hip/.hiplc, .aep). Common cause: dragging only the textures folder instead of the full project root. Re-upload from one level up so the scene file is included.

"Asset paths point outside the project folder."

Your scene references assets that live above the project root (e.g., a shared \Studio Library\ folder one level up from \Project_X\). Worker nodes can only see what's inside the uploaded folder. Either consolidate those assets into the project folder locally before upload (most DCCs have a "collect all assets" command), or upload the parent folder so all referenced paths resolve.

Downloading Rendered Output

Two paths exist for retrieving completed frames: manual download from the dashboard, or automatic streaming via the Client App.

Manual download via dashboard. When a job completes, the dashboard lists each output frame and offers per-frame and full-job download. For animation jobs, the dashboard zips the output (this is a download convenience โ€” not a submission format) so you can pull the whole sequence in one transfer. For stills or short sequences, per-frame download keeps the file count manageable.

Auto-download via Client App. With the Client App installed and a local output folder configured, completed frames stream down as each frame finishes. You don't have to wait for the full job to complete โ€” you can review frame 1 while frames 240 through 480 are still rendering. This is especially useful for long animation jobs where review cycles overlap with the render itself.

Retention. Output files stay available for download after the job completes; there is no fixed automatic deletion period, and we delete a project when you ask us to. For long-term archival, download to your studio's storage rather than relying on the farm as a backup tier.

FAQ

Q: Why doesn't Super Renders Farm extract ZIP files? A: The render manager needs to walk your project folder to resolve asset paths: scene files reference textures, proxies, caches, and plugins by relative or absolute paths. The farm does not unpack archives of any format, so upload the project as an uncompressed folder, which is what every DCC's "collect all assets" command produces.

Q: What's the largest project I can upload? A: There's no enforced upper limit. We've handled multi-hundred-gigabyte animation projects with cached simulations. For large projects use the Client App: browser uploads stall on long transfers, and the Client App resumes where it left off.

Q: How do I upload if my connection is unreliable? A: Use the SuperRenders Client App. It resumes on disconnect, so a dropped Wi-Fi connection doesn't restart your transfer from zero. The browser upload path doesn't resume: if your tab closes mid-upload, the transfer aborts.

Q: How long are my rendered frames kept on the farm? A: Output files stay available for download after the job completes; there is no fixed automatic deletion period, and we delete a project when you ask us to. For archival, download completed jobs to your studio's storage rather than treating the farm as long-term backup.

Q: How do I get completed frames back to my workstation automatically? A: Use the SuperRenders Client App with a configured output folder. Frames stream down as each one finishes, so you can review early frames while later frames are still rendering. The dashboard's manual download path is fine for one-off jobs, but for animation work the Client App's auto-download cuts a lot of waiting.

Q: My scene references textures on a network drive. Will those upload? A: Only if the network drive paths are inside your uploaded project folder. Worker nodes can't reach your studio's network shares directly. Run your DCC's "collect all assets" or "consolidate references" command locally first โ€” that copies all referenced files into the project folder and rewrites the paths to relative โ€” then upload the consolidated folder. Most failed-job tickets we see trace back to this step being skipped.

Formats and size limits, in depth: why the farm needs your project unpacked and the ~300 GB single-upload guideline are covered in the dedicated guide: Upload formats and size limits.

Last updated: September 29, 2026