September 2026 LiveRamp Clean Room Release Notes
Product Updates – General Availability
External API Support for Clean Room Templates
You can now manage Clean Room templates through the Clean Room API, so automation-heavy teams can scale their template workflows without relying on the UI. With the API, you can:
Create and edit templates: Build and update Clean Room templates programmatically.
Generate clones: Create clean rooms from a template.
Add template items to existing clean rooms: Add questions and flows from a template to a clean room that already exists.
For more information, see "List cleanroom templates".
Historical Run Exports for Flows
Building on Historical Run Exports for Questions, you can now add export jobs to flow nodes and trigger exports for runs that occurred up to 90 days in the past. This helps if you forgot to set up an export before running a flow, need the results for BI or QA purposes, or need to re-send a previous job after a data ingestion issue on your end.
For more information, see "The Flow Runs Page".
RampID and Known ID API Support
The external API now supports identity resolution workflows for clean rooms. With the API, you can:
Mark a data connection for ID resolution: Include a schedule and country.
Review mapping file status: Check the status of your mapping files.
Use resolved mapping files in clean rooms: Provision RampID and Known ID mapping files and assign them to questions.
For more information, see "Map Field Configurations with Identity Resolution".
Hashed MAID and Hashed Phone Number Support for International Identity Resolution
International customers can now use hashed MAIDs and hashed phone numbers for RampID resolution via Snowflake.
For more information, see "Getting Started with Identity Resolution in LiveRamp Clean Room (Non-US Data)".
Auto-Scale Compute Capacity
For Hybrid clean rooms, we’ve added an experimental feature that allows you to set the compute capacity of a question or flow node to Auto-Scale. The system automatically tunes run configurations based on the question's code and data volume, and those configurations persist for future runs.
With Auto-Scale enabled:
The next run of a question will generally use settings roughly equivalent to ~medium compute, though it will adjust for particularly large queries/data.
If that run fails with the initial settings, the run will retry automatically (no failure will be surfaced in the UI).
On retry, the system will evaluate the specific cause of the failure (such as out of memory or CPU limits) and upgrade that particular setting. It will then restart the run.
The cycle will repeat until the run succeeds. At that point, the settings are saved and will be reused for the next run.
Any edits to the question that change the capacity needs might cause the cycle to restart.
To change a question’s compute capacity to Auto-Scale:
Enter the desired clean room.
Go to the Questions page.
In the row for the desired question, click the More Options menu (the three dots) and select Advanced Settings.
Select (Experimental) Auto-Scale and then click .
You will likely see a longer-than-usual runtime for the first run while the system is calibrating. Once the settings are dialed in, we expect runs for that question to result in fewer failures and manual adjustments.
Dataset Assignments Retained When Editing Questions
Previously, editing a question unassigned all datasets so partners could review the change. Now, datasets owned by the question author's organization stay assigned after an edit, as do partner datasets with the Partners Can Assign setting enabled. This makes iterating on a question faster by reducing the number of datasets you need to reassign.
For more information, see "Edit a Question".
Product Updates – Limited Release
Semantic Context and Multi-Layered Semantic Data Endpoints in the External API
New create, read, update, and delete endpoints (Clean Room Contexts) are available in the external API, so agentic tools can read and write semantic context about a clean room.
You can use these endpoints in the following ways:
Clean room purpose and groupings: Describe what the clean room is for and how its questions and flows are organized.
KPI definitions: Define the key metrics used in the clean room.
Collaborative intelligence context: Add context defined by owners and partners.
This context is available through the API and MCP instructions.
For more information, see "Get clean room context".
Quick Hits and Fixes
User groups in templates: You can now add a user group to a template clean room, and the group is automatically invited to any clean room cloned from that template. For more information, see “Add Users to a Clean Room”.