Pipelines
Automated Render Workflows
Beta: The Glossi API is currently in beta. Endpoints, request/response formats, and behavior may change as we iterate. If you run into issues or have feedback, reach out at [email protected].
Pipelines connect an external file storage system — like Google Drive or OneDrive — directly to Glossi. When new 3D model files appear in your source folder, Glossi automatically picks them up, processes them, renders images using your chosen template, and exports the finished renders to your output folder.
No manual steps — just drop files and get renders.
How Pipelines Work
A pipeline watches your source folder for new 3D model files. For each file it finds, it runs through a fully automated workflow:
Queued
New file detected in source folder
Ingesting
File downloaded and uploaded to Glossi
Processing
3D model converted and prepared
Rendering
Images and videos rendered using your template
Styling
Brand Style variants generated (if styles are configured)
Exporting
Finished renders uploaded to your output folder
Completed
Done
If any stage fails, the pipeline retries automatically based on your retry settings.
Creating a Pipeline
Go to Pipelines in the navigation bar and click "Create Pipeline". The guided form walks you through five steps:
1. Connectors
Choose your storage provider and sign in to connect your account. Currently supported: Google Drive and OneDrive.
Input source — The folder Glossi watches for new model files.
Output destination — The folder where finished renders are exported.
You can use the same account for both input and output, or connect separate accounts. Use the folder picker to browse and select your folders.
2. Basics
Pipeline name — Give your pipeline a descriptive name (e.g., "Product Renders", "CAD to Images").
Template — Choose the template that defines lighting, camera angles, and environment for all renders.
3. Render Settings
Configure the quality of your rendered output:
Render bookmarks — Render all camera bookmark images from the template.
Render shots — Render all video shots from the template.
Render variants — Render all model variants if the model has them.
Image quality — 720p, 1080p, or 4K resolution.
Video quality — Fastest (MP4), Balanced, or Best (ProRes).
4. Brand Styles
Optionally attach Brand Styles to your pipeline. When styles are configured, Glossi automatically generates styled variants for every image render — one variant per selected style.
For example, if your template has 3 camera bookmarks and you select 2 styles, each model will produce 3 base renders + 6 styled variants (3 renders x 2 styles).
Group by style — Controls how styled variants are organized in your output folder:
Off (default) — Styled variants sit alongside base renders with the style name appended to the filename (e.g.,
Front View_Warm Studio.png).On — Styled variants are organized into style-named subfolders (e.g.,
Warm Studio/Front View.png).
Tip: You can add or remove styles from an existing pipeline at any time from the pipeline detail page. Changes apply to future jobs only — renders already in progress continue with their original style configuration.
5. Advanced Settings
These defaults work well for most pipelines:
Poll interval
5 minutes
How often Glossi checks your folder for new files
Batch size
10
Maximum files to process per cycle
Max retries
3
How many times to retry a failed file before giving up
Reprocess modified files
Off
Re-render a file if it's updated after its initial render
Process existing files
On
Render files already in the folder when the pipeline is first created
Monitoring Your Pipeline
Click on any pipeline card to open its detail page. Here you can see:
Job Status Overview
A status grid shows the count of jobs in each stage — Queued, Ingesting, Processing, Rendering, Styling, Exporting, Completed, and Failed. This gives you a quick snapshot of pipeline activity.
Job List
Browse individual jobs with filtering by status. Each job shows:
File name — The source model file
Status — Current processing stage
Outputs — Number of rendered files
Error — Error message if the job failed or had partial failures
Pipeline Configuration
View and edit all pipeline settings directly from the detail page, including:
Render settings (quality, bookmarks, shots, variants)
Brand Styles (add, remove, toggle grouping)
Connector details and reconnection if auth expires
Managing Jobs
Delete & Re-run
Select one or more jobs and click "Delete & Re-run" to remove them from the pipeline's tracking. The files will be re-detected on the next poll cycle and processed from scratch.
Retry Styling
If style variants fail for a job (due to a temporary AI service issue, for example), you can retry just the styling step without re-rendering. Select the affected jobs and click "Retry Styling" — the base renders are preserved and only the failed styled variants are regenerated.
Jobs with partial styling failures show an amber warning icon next to their status.
Managing Your Pipeline
From the pipeline detail page, you can:
Run Now — Trigger an immediate poll of the source folder instead of waiting for the next scheduled poll.
Pause / Resume — Temporarily stop or restart the pipeline.
Delete — Remove the pipeline and all its job history.
Output Folder Structure
Rendered files are organized in your output folder by model name:
When Brand Styles are attached with Group by style enabled, styled variants are organized into subfolders:
With Group by style off (default), styled variants sit alongside base renders:
How File Detection Works
Pipelines use efficient change tracking to detect new files:
First poll — Scans the folder and either queues all existing files (if Process existing files is on) or just starts tracking from that point.
Subsequent polls — Only checks for changes since the last poll, so even a folder with thousands of files is scanned instantly if nothing changed.
Deduplication — Each file is only processed once. If the same file appears in multiple scans, it's ignored after the first time.
Modified files — If Reprocess modified files is enabled, files that are updated after their initial render will be re-processed.
Error Handling
Pipelines are designed to handle failures gracefully:
Automatic retries — Failed files are retried with increasing delays (1 minute, 2 minutes, 4 minutes, and so on).
Max retries — After the configured number of retries (default 3), the file is marked as permanently failed.
Circuit breaker — If 3 consecutive processing cycles make zero progress (e.g., expired auth or a misconfigured connector), the pipeline pauses automatically. Resume it from the dashboard once you've investigated.
Job history — Every failure is recorded with the error message, the stage it failed at, and a timestamp.
Supported File Types
.glb
glTF Binary
.gltf
glTF
.usdz
Universal Scene Description
.usd
Universal Scene Description
.fbx
Autodesk FBX
.obj
Wavefront OBJ
.step, .stp
STEP (CAD)
.stl
Stereolithography
Available Connectors
OneDrive
Available — personal and Microsoft 365 business accounts
Google Drive
Available — personal and Google Workspace accounts
Dropbox
Coming soon
Amazon S3
Coming soon
Last updated
Was this helpful?