Issue Station
Sign In
Tools hereToolbox

hermes - 💡(How to fix) Fix [Feature]: Analytics — separate Input/Output token charts + per-model breakdown [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
NousResearch/hermes-agent#20412•Fetched 2026-05-06 06:36:54
View on GitHub
Comments
0
Participants
1
Timeline
4
Reactions
0
Author
sanchomuzaxsanchomuzax
Participants
sanchomuzaxsanchomuzax
Timeline (top)
labeled ×3subscribed ×1

The current Daily Token Usage chart on the Analytics page (/analytics) displays Input and Output token counts on the same graph. Because Input tokens are typically an order of magnitude larger than Output tokens, the Output line is nearly invisible and the chart is hard to read at a glance.

Root Cause

The current single merged chart hides actionable signal:

  • Input spikes are often caused by large context windows / long conversation chains — seeing them separately helps tune chunking/summarization strategies.
  • Output spikes indicate verbose responses — useful for prompt tuning.
  • The ratio trend reveals whether the agent is becoming more or less context-efficient over time.

On behalf of Sancho, by a Hermes Agent

RAW_BUFFERClick to expand / collapse

Summary

The current Daily Token Usage chart on the Analytics page (/analytics) displays Input and Output token counts on the same graph. Because Input tokens are typically an order of magnitude larger than Output tokens, the Output line is nearly invisible and the chart is hard to read at a glance.

Proposed Changes

1. Split Input and Output into separate panels

Three vertically stacked panels would immediately surface the relationship:

PanelContent
Input tokens/dayTime-series line chart in millions
Output tokens/dayTime-series line chart in thousands
Input/Output ratioHow many times more input than output per day, with average line

This makes it easy to answer questions like:

  • On which days did context pressure peak? (Input spike)
  • On which days were responses especially long? (Output spike)
  • Is my agent spending more time reading or generating? (Ratio trend)

The chart below was generated from 31 days of real usage data (Apr 5 – May 5, 2026) and illustrates what this would look like:

<img width="2084" height="1769" alt="Image" src="https://github.com/user-attachments/assets/c7570d67-b0d7-442e-9a7b-eaae18f11101" />

2. Per-model token breakdown (similar to OpenRouter)

OpenRouter's usage dashboard shows a cumulative stacked area/bar chart of token consumption broken down by model over the selected time period. Something similar would be very valuable in Hermes Agent Analytics:

  • Which models are responsible for the highest token consumption?
  • How has model usage distribution shifted over time? (e.g. after a config change)
  • Which model dominates Input vs. Output costs?

A stacked bar chart per day (or a cumulative area chart for the full period) with one color per model would make this instantly readable.

Why This Matters

The current single merged chart hides actionable signal:

  • Input spikes are often caused by large context windows / long conversation chains — seeing them separately helps tune chunking/summarization strategies.
  • Output spikes indicate verbose responses — useful for prompt tuning.
  • The ratio trend reveals whether the agent is becoming more or less context-efficient over time.

On behalf of Sancho, by a Hermes Agent

extent analysis

TL;DR

Split the Daily Token Usage chart into separate panels for Input and Output tokens to improve readability and surface actionable insights.

Guidance

  • Consider implementing three vertically stacked panels: one for Input tokens, one for Output tokens, and one for the Input/Output ratio, to provide a clear visual representation of the data.
  • Use a time-series line chart in millions for Input tokens and in thousands for Output tokens to account for the difference in scale.
  • Include an average line in the Input/Output ratio panel to help identify trends over time.
  • Evaluate the feasibility of adding a per-model token breakdown, similar to OpenRouter's usage dashboard, to provide further insights into model usage and token consumption.

Example

No code snippet is provided as the issue focuses on UI/UX changes and data visualization rather than code implementation.

Notes

The proposed changes aim to address the issue of the current chart being hard to read due to the large difference in scale between Input and Output tokens. By separating the data into distinct panels, the new design should provide a clearer and more actionable visualization of the data.

Recommendation

Apply the proposed workaround of splitting the Daily Token Usage chart into separate panels for Input and Output tokens, as it directly addresses the issue and provides a clear path to improving the readability and usefulness of the chart.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#ssr#memory leak#API versioning#request timeout

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING