← Back to glossary
+Suggest a term
Concept·Patterns & Practices·Added 1 month ago

Prompt versioning

Also known as: prompt version control, prompt management, prompt registry

Treating prompts like code: tracking changes over time, running evals against each version, and maintaining the ability to roll back when a change degrades performance. A basic hygiene practice for any AI system in production.

Early AI projects often live in a single file or a shared doc where prompts get edited in place. When the system behaves worse, nobody knows what changed or when. Prompt versioning applies the same version control discipline used for code to the prompts that drive AI behavior.

In practice this means storing prompts in version-controlled files (tracked in Git, the most common code version control system), tagging releases, and running your eval suite against each version before promoting changes. If a new prompt version hurts performance on your test cases, you can see exactly what changed and revert.

Some teams use dedicated prompt management tools like Langfuse or PromptLayer that layer dashboards and experiment tracking on top of version control. For smaller projects, plain Git and a disciplined naming convention gets you most of the way there. Either way, the core behavior change is the same: stop treating prompts as informal text and start treating them as software artifacts with the same care you'd give to code.

This definition is AI-generated and refreshed weekly. It may contain inaccuracies. Use your own judgment, especially for production decisions.
Related terms
Prompt engineeringEvalsSystem promptLLMOpsSpec-driven development