You can now run AutoReview against your company's own checklist templates, giving you more control over the checks AutoReview performs.

How it works

There are two steps. One for admins, one for all users.

Admins: Go to Company Settings → General → Checklist Templates. Each template now has an "Available in AutoReview" toggle.

Enabling a template runs a safety check on the content before it becomes available. Templates that pass will appear for all users across all workspaces. Templates that fail the check will be marked incompatible and won't appear in AutoReview runs.

All users: When starting a Peer Checker run, open the checklist selector. Your company's enabled templates appear under Your Checklists, separate from CoLab's built-in options. Select one and run as normal.

‍

What to expect from results

AutoReview uses your checklist as input to guide its analysis — it passes the checklist name, items, and descriptions to its underlying agents. However, results are still limited by the check types those agents support. If your checklist includes items outside current agent coverage, AutoReview may not surface findings for those items. This scope will expand as new agents are added.

If you're getting fewer results than expected, this is the most likely reason. Your CSM can help you identify which checklist items are most likely to produce results with the current agent set.

‍

A few things to know

• Only company admins can enable or disable templates for AutoReview
• Enabled templates are available across all workspaces
• You can still use CoLab's built-in checklists at any time

For more informaton, read our Knowledge Base article.

AutoReview for 2D drawings now runs a broader set of specialized agents, significantly expanding the range of technical feedback it can generate.

Previously, 2D AutoReview used a single generalist agent and primarily surfaced feedback on missing dimensions, material inconsistencies, and text typos. AutoReview now dynamically selects from a factory of specialized agents based on the content of the drawing being reviewed — covering GD&T, tolerancing, dimensioning, fasteners, view integrity, BOM consistency, and more.

Custom checklists and user input prompts work alongside the agent factory — they direct which agents are prioritized and what context they operate with. The active agents for each run are visible in the processing panel.

If you've run AutoReview on a drawing previously, you will likely see new findings on the same file. This reflects expanded coverage, not a change in the design. The agent factory will continue to expand over time.

Import feedback via CSV

You can now upload a CSV file to bring existing design feedback into CoLab and attach it to a specific file.

To import feedback:

• Open the file the feedback should be associated with
• Select Import feedback via CSV from the Feedback menu
• Choose a feedback type for the imported feedback items
• Upload your CSV file
• Add feedback pins and relevant markups

Each row in the CSV becomes a feedback item on that design.

Once imported, the feedback behaves like feedback created directly in CoLab. It can be reviewed, discussed, and tracked in the same way.

Learn more about importing feedback from outside of CoLab here.

‍

New Portal Guest permissions for review creation and management

‍

Admins can grant portal guests permissions that let them:

• Create reviews on files they upload
• Edit review details
• Act as review owners
• Close reviews
• Export review history

While these new options let external collaborators initiate and manage reviews, admins maintain full control what portal guests can and cannot do.

CoLab now offers a versioned REST API that lets you access and manage CoLab data and actions programmatically. You can use the API to perform bulk actions in the application, and pull data from CoLab into other tools.

Admins generate API keys in Admin Settings > API Key Management and use those keys to authenticate requests to the Enterprise API. Once authenticated, you can:

• Add, update, or remove users from workspaces
• Create and manage workspaces
• Upload, update, and delete files in bulk
• Retrieve reviews and feedback for use in BI tools, databases, or IT workflows

More information on generating API keys can be found here.

API actions follow the same permission rules as the CoLab UI, and changes appear immediately across the product.

For more information, see our knowledge base or contact your Customer Success Manager.

‍

‍

You can now run wall thickness analysis directly on 3D models in CoLab.

This allows you to evaluate material thickness during a review without leaving CoLab or reopening CAD. Thickness variation is displayed on the model using a color map, making thin sections, thick transitions, and other potential manufacturability concerns easy to spot.

• Thickness is calculated across BREP geometry and visualized directly on the model
• The color scale reflects min/max thickness values and can be adjusted to focus on specific ranges
• A probe tool lets you inspect exact thickness values at any point on the model
• Feedback created during analysis preserves the thickness view and scale, so the context is retained when revisiting the issue later

This capability supports decisions around whether geometry is ready to move forward, what needs adjustment, and what should be flagged ahead of tooling or supplier review.

Wall Thickness Analysis is available on Pro and Scale plans.

Read more in our knowledge base article.

You can now provide AutoReview with context about a design before it runs — so the analysis is informed by information that isn't captured in the file itself.

When triggering an AutoReview run, an optional input field lets you describe the design in plain text. AutoReview uses this context when generating feedback across both general runs and checklist-based runs. Useful context to include:

• Material or material grade (e.g., PA66 with 30% glass fill)
• Intended fabrication method (e.g., CNC machining, injection molding, sheet metal)
• Application details (e.g., operating environment, load conditions)
• Specific instructions for what AutoReview should evaluate

The prompt is saved in the run header so it's visible to anyone reviewing the results later. Leaving the field blank runs AutoReview as normal.

Note: AutoReview no longer runs general analysis automatically on file upload by default. This can be re-enabled in Company Settings → AutoReview.

AutoReview now shows what it's doing while it processes, so you can follow along rather than wait for results to appear.

When you trigger a run, a processing panel shows the active stages in sequence — which agents are running, what they're analyzing, and where the run stands overall. Stages update in real time as the run progresses.

This replaces a blank waiting state that gave no indication of progress. The intent is to make AutoReview's process legible enough that the results make sense when they arrive — and to give users more confidence in what the AI is doing and why.

Processing stages are shown during active runs only — completed runs from before this launch won't show them. The set of visible stages will expand as more agents are added.