LLMTrace Dashboard

Sessions

Session Name:

User ID:

Model Name:

Message Content:

Start Date (YYYY-MM-DD):

End Date (YYYY-MM-DD):

Min Tokens:

Max Tokens:

Min Cost:

Max Cost:

Session Details:

Start:

End:

Total Tokens:

Total Cost: $

User ID:

Conversation

Metrics

Feedback

Errors

Aggregated Metrics

Average Session Duration: N/A seconds

Metrics Time Series

Select Metric:

Time Interval:

Start Date:

End Date:

Model Token Usage

Feedback Score Distribution

Error Type Counts

About LLMTrace

LLMTrace is an open-source Python library for LLM application observability and evaluation. It helps you track, log, and analyze interactions with Large Language Models.

Key Features:

Automatic instrumentation for popular LLM frameworks (OpenAI, LangChain, HuggingFace).
Session-based logging of prompts, responses, tokens, costs, metrics, feedback, and errors.
SQLite database for local, persistent storage.
Programmatic API for data retrieval.
Web-based dashboard for visualization and analysis.
CLI for common tasks like listing sessions, showing details, and exporting data.

For more information, visit the project's GitHub repository.