Prompt management.Ready for production.
Give every prompt a version history and every release a deliberate decision. Edit, evaluate, publish, and retrieve your prompts in one connected workflow.
No credit card required. Product access is subject to early-access approval.
A shorter instruction
CandidateSystem message
You are a support triage assistant. Categorize each ticket as billing, technical, urgent, or general.
Reply with the single most specific category.
Version 2 removes the escalation and output-format rules. Evaluate the change before moving production.
Sample data · Illustrative workflow, no live model calls.
PromptLens Prompts is a prompt management product for engineering teams building LLM features. It brings personal drafts, immutable versions, evaluation evidence, and publishing labels together around each named prompt.
From a first draft to your next release.
Keep the prompt, its test cases, and its release history together. Every step has a clear place in the workflow.
Edit
Work in a personal, autosaved draft with messages, variables, and model settings.
Version
Save an immutable revision so you can return to the exact prompt you tested.
Evaluate
Test the saved version on your prompt dataset and compare it with production.
Publish
Move production, staging, or a custom label to the version you choose.
Retrieve
Load a version by label or exact number through the prompt retrieval API.
A shared history. Room to experiment.
Keep unfinished edits in your personal draft. Save a version when you want a permanent record of its messages, template variables, and model settings.
Personal drafts · Immutable versions
Make the release decision explicit.
Review evaluation evidence before moving a label. If you need to roll back, point production to an earlier saved version. The next retrieval request sees that assignment.
Production and staging · Custom labels · Rollback
Your application gets the version you choose.
Retrieve a saved prompt by its permanent ID using a key scoped to that prompt. Labels let you change the selected version without redeploying the prompt text in your application.
Read-only retrieval API · Exact versions or labels
Version the change. Then test the difference.
A prompt diff shows what you edited. Evaluations show how that edit changes the answers. Use both before deciding what becomes production.
Explore EvaluationsQuestions about prompts.
Straight answers about prompts, how it works, and where to start.
Compare plans and limits- What is prompt management?
- Prompt management is the practice of organizing, versioning, testing, and publishing the prompts used by an AI application. PromptLens keeps each named prompt's saved versions, dataset, and publishing labels together so a team can review changes before using them in production.
- How does prompt version control work?
- Edits are autosaved in your personal draft. Saving creates an immutable prompt version containing its messages, template variables, and model settings. Your team can inspect saved versions and assign labels to the version it wants to use.
- Can I roll back a production prompt?
- Yes. Move the production label to an earlier saved version. The next request to the retrieval API for that label returns the newly assigned version. An application using an exact version number remains pinned to that version.
- Does the retrieval API call the model?
- No. It returns saved prompt messages and settings. Your application handles variable rendering and model execution. Authenticate with an integration key scoped to one prompt, and select either a label or an exact version.
- Do I need an evaluation before publishing my first prompt?
- Initial prompt setup requires a completed, fully scored baseline evaluation. Some test cases may fail; there is no minimum pass rate for completing setup. The initial saved version establishes staging and production.
- Can I start with the Free plan?
- Yes. Free includes up to 3 prompts, 2 organization members, and 10 test cases per prompt. Signup follows the existing early-access approval flow. See Pricing for current plan limits and model usage details.
Give your next prompt change a clear path to production.
Start with a prompt, add your test cases, and establish your first baseline.