Prompt
I need prompt version comparison so I can see what changed after a deploy
Latest observation
Oct 9, 2026
OpenAI APIWeb search: off
Sure — if you want to compare prompt versions after a deploy, the cleanest way is to compare:
- version metadata
- the actual prompt text
- parameters/settings
- attached tools / system instructions
- runtime behavior logs, if available
Quick checklist for prompt version comparison
Compare these fields
- Version ID / commit hash
- Created at / deployed at
- Prompt content
- system prompt
- developer prompt
- user template
- Variables / placeholders
- Model settings
- temperature
- top_p
- max tokens
- Tool/function definitions
- Safety / policy text
- Routing logic
- Few-shot examples
Recommended diff format
Use a side-by-side or unified diff like:
- You are a helpful assistant.
+ You are a concise helpful assistant for internal support.
- Always answer in detail.
+ Answer briefly unless the user asks for more detail.
If you want a deploy comparison report
A useful report usually includes:
- What changed
- Why it matters
- Risk level
- Expected behavioral impact
- Rollback recommendation
Example:
| Field | Before | After | Impact |
|---|---|---|---|
| System prompt | Generic assistant | Support-focused assistant | Tone changed |
| Temperature | 0.7 | 0.2 | More deterministic responses |
| Tool access | None | Search tool enabled | New external behavior |
If you need this automated
You can compare:
- prompt store version
AvsB - deploy revision
oldvsnew - exported JSON/YAML prompt configs
If you want, I can help you with any of these:
- a prompt diff template
- a JSON schema for version comparison
- a script to diff two prompt versions
- a change report format for deploys
If you paste the two prompt versions, I can compare them directly.