traceloop / traceloop/openllmetry
🚀 Feature: Option to detect and visualize repetitive text blocks
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 7.4k
- Forks
- 1.1k
- Avg merge
- 8d 14h
- Merged PRs (30d)
- 2
Description
Which component is this feature for?
All
🔖 Feature description
Describe the feature or potential improvement
When tracing agents, it's common to see prompts that combine a) variable text and b) static text defined in templates that are included in multiple prompts (for example, a system prompt).
It would be nice if TraceLoop cuold identify repetitive text patterns (prompt templates) and visualize them in other color, inside a text container that can be labeled, expanded and collapsed.
Example
Let's say we traced this input prompt:
We can now click the text block and assign it a custom label:
When viewing other prompts that contain the same text block, it will be displayed using our custom label:
This feature will make it much more easier to analyze prompts.
Optional
Please also consider that some templates have variables inside.
Eg: "The current date is {{date}} and the current time is {{time}}."
In this case, we would see this:
It would be nice if we could tell TraceLoop to treat it as just one block that contains variables.
This way we could add it a custom label and only see:
Additional information
No response
🎤 Why is this feature needed ?
To make it easier to trace and analyze prompts.
👀 Have you spent some time to check if this feature request has been raised before?
- I checked and didn't find similar issue
Are you willing to submit PR?
None
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue names no files, tests, or entry points; begin by locating the TraceLoop prompt-trace visualization and reviewing how prompt text is represented. Clarify matching, variable handling, labeling, and collapse behavior before implementation; done means repeated blocks are consistently detected and visualized across prompts.
Written by the indexing model from the issue text.
Assessment
- Domain
- ai, observability
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100