What does the model and prompt version register produce?
Link a deployed model, prompt or configuration version to its owner, change summary, evaluation evidence and approval state.
Prepare and review the task-specific records in this browser. Inspect consistency findings and retain the complete backup before preparing the finalized pack. Accountable decisions require responsible review.
What do you need before you start?
- Define the named scope, accountable owner and your review basis.
- Retain the actual evidence separately. Recorded links and evidence references are declarations and are not fetched.
- Choose a valid review date. Resolve missing fields, contradictory declarations and date or status findings before relying on the review.
What information do you enter?
The browser workspace asks only for the structured fields listed below.
- AI system (required). Use the name in the organization’s inventory.
- Versioned component (required). Choose the exact component that changed.
- Provider or maintainer (required). Name the external provider or internal maintaining team.
- Model or component version (required). Use the exact available identifier.
- Prompt or configuration version (required). Use the organization’s controlled version identifier.
- Deployment or release date (required). Record the customer-reported date.
- Change summary (required). Describe what changed and why.
- Evaluation evidence reference (required). Link the version to an internal test or review record.
- Release owner (required). Name the person or role responsible for the release.
- Approver (optional). Name the person or role that approved the release.
- Release status (required). Record the customer-reported state.
| Field | What to enter | Fictional example |
|---|---|---|
| AI systemRequired | Use the name in the organization’s inventory. | Example policy assistant, fictional |
| Versioned componentRequired | Choose the exact component that changed.Choose from Model, System prompt, User prompt template, Retrieval configuration, Tool configuration, Safety configuration. | System prompt |
| Provider or maintainerRequired | Name the external provider or internal maintaining team. | Internal product team |
| Model or component versionRequired | Use the exact available identifier. | example-model-2 |
| Prompt or configuration versionRequired | Use the organization’s controlled version identifier. | prompt-2026-09-01 |
| Deployment or release dateRequired | Record the customer-reported date. | 2026-09-10 |
| Change summaryRequired | Describe what changed and why. | Added a required source citation instruction and a refusal boundary. |
| Evaluation evidence referenceRequired | Link the version to an internal test or review record. | EXAMPLE-EVAL-021 |
| Release ownerRequired | Name the person or role responsible for the release. | AI product owner |
| ApproverOptional | Name the person or role that approved the release. | AI governance lead |
| Release statusRequired | Record the customer-reported state.Choose from Draft, Approved, Deployed, Rolled back, Retired. | Draft |
What you can download
Every result must identify its supplied facts, assumptions and unresolved items.
| Free result | Included downloads |
|---|---|
Enter your information, inspect the on-screen result and keep a browser backup. Editing and direct copies of your entered information stay free. |
These supported outputs are free. No checkout is required. |
How does the model and prompt version register work?
- 1Prepare the named records
Use the task-specific fields above. Retain original evidence separately and record exact references and versions.
- 2Review internal consistency
Choose the review date. Inspect each scope identity, contradictory declaration, missing decision support and date or status finding.
- 3Keep recoverable work
Download a free backup before replacing or clearing work. Restore earlier record backups or the complete current snapshot, including unfinished drafts.
- 4Prepare the finalized pack
Prepare all nine files, then download them free. Review excerpts in PDF and DOCX against the complete Excel, CSV and JSON. Optional private cloud saving requires sign-in and separate consent.
Method and sources
Apply the steps above to the supported inputs. Use the listed source sections to review the rule basis, assumptions and limits.
- Artificial Intelligence Risk Management Framework 1.0A voluntary US baseline for AI risk records, responsibilities, evaluation and lifecycle decisions. Govern, Map, Measure and Manage.National Institute of Standards and Technology. Published 2023-01-26. Reviewed 2026-10-08.
- AI Risk Management Framework PlaybookImplementation ideas and documentation practices for the AI RMF. Suggested actions for Govern, Map, Measure and Manage.National Institute of Standards and Technology. Published 2023-03-30. Reviewed 2026-10-08.
Common questions
What information does the model and prompt version register need?
Enter AI system (required). Use the name in the organization’s inventory; versioned component (required). Choose the exact component that changed; provider or maintainer (required). Name the external provider or internal maintaining team; model or component version (required). Use the exact available identifier; prompt or configuration version (required). Use the organization’s controlled version identifier. Define the named scope, accountable owner and your review basis.
What does the model and prompt version register produce?
Complete all-field record review and task-specific scope rows. Every unresolved finding stays visible. Free restorable browser backup including the unfinished draft, editing reference and review date. Nine-file finalized pack. Formal PDF, editable DOCX, complete Excel register and review sheets, records CSV, findings CSV, scope review CSV, complete JSON, file hashes and instructions. Review unresolved findings and limitations before use.
Who makes the final decision?
A responsible person reviews the result. Checks review customer declarations, recorded dates and internal consistency. They do not authenticate evidence, approve an AI use, determine legal applicability or establish conformity, safety or risk acceptance.
Is this tool free?
Yes. The on-screen result and listed downloads are free. No checkout is required. Keep local backups and review the stated method and limits.
Can I use the model and prompt version register for this task?
Link releases to configurations and test evidence. Checks review customer declarations, recorded dates and internal consistency. They do not authenticate evidence, approve an AI use, determine legal applicability or establish conformity, safety or risk acceptance.
What still needs review?
Checks review customer declarations, recorded dates and internal consistency. They do not authenticate evidence, approve an AI use, determine legal applicability or establish conformity, safety or risk acceptance.
Each workspace supports 1000 records, 20000 characters per field and a 30 MB complete workspace. Invalid and foreign backups preserve current browser work.
PDF and DOCX show the first 20 records, field excerpts through 2000 characters and the first 40 findings. Excel, CSV and JSON retain complete values and derived rows. Referenced files are not fetched or included.