Hermes Telemetry
nujovich/hermes-telemetry
Budget enforcement and observability plugin for Hermes Agent that stops runs before they overspend
Hermes Telemetry is a plugin for Hermes Agent that tracks token usage and cost in real time and enforces spending limits before the next model call. It adds /stats and /budget commands for reviewing spend and setting caps.
What Hermes Telemetry does
Hermes Telemetry runs inside the Hermes runtime through hooks, so it sees every API call across sessions, platforms and cron jobs. It records tokens and cost to SQLite in WAL mode, syncs OpenRouter pricing automatically, and checks budget limits before the next LLM call is made. That lets it block work that would overspend, such as a scheduled cron run, instead of only reporting the cost afterward.
The user-facing surface is two commands: /stats for usage and cost breakdowns, including per-model views, and /budget for viewing and setting caps. For example, /budget set global daily 0.001 writes a daily cap to budget.yaml. Prices can be overridden in pricing.yaml, and a model id ending in :free is treated as $0 and flagged for a free-to-paid alert. The README says it targets the gap around budget enforcement for unattended runs raised in the Hermes issue tracker. It was built for the Hermes Agent Challenge by Nadia Ujovich.
Key features
- Real-time token and cost tracking through Hermes runtime hooks
- Budget limits checked before the next LLM call
- /stats and /budget commands, including per-model stats
- Budget caps in budget.yaml and price overrides in pricing.yaml
- Automatic OpenRouter pricing sync with SQLite logging in WAL mode
- Free-tier handling for model ids ending in :free
When to use it
- Capping daily spend for unattended cron jobs
- Finding which models use the most tokens and money
- Getting an alert when a free-tier model starts billing
Who it is for: Hermes Agent users who run autonomous, scheduled or multi-platform agents and need hard spending limits.
How it fits with Hermes Agent
Built as a Hermes plugin that hooks into the Hermes runtime and adds the /stats and /budget commands.
FAQ
What is Hermes Telemetry?
Hermes Telemetry is a Hermes Agent plugin for cost tracking and budget enforcement. It logs tokens and spend, and can stop work that would exceed a configured cap.
Can Hermes Telemetry stop a run that is overspending?
Yes. It enforces budget limits before the next LLM call, so a run or cron job that would exceed the cap is blocked rather than only reported afterward.
Is Hermes Telemetry free and open source?
Yes. The repository is released under the MIT license.
Similar models for Hermes Agent
All modelsOpenAI-compatible proxy that rotates free-tier LLM accounts with per-model quota routing
piyush-tyagi-13 llm-keypoolFree-tier LLM API key pool with rotation, 429 cooldowns and a local OpenAI-compatible proxy
InfiniteWhispers HermesAgent-MultiModelFully local multi-model setup for Hermes Agent with Ollama, Mixture-of-Agents routing and a tuning guide
open-world-project model-routerHermes Agent plugin that routes each turn to the cheapest of five model tiers and escalates when needed
KaiFelixBennett Hermes Local StackRun Hermes Agent and Claude Code on a local llama.cpp model with no API costs
compnew2006 MetaAI Free Hermes AgentSelf-hosted Go API that wraps Meta AI's Muse Spark as an OpenAI-compatible endpoint for Hermes Agent
Related guides: How to install Hermes Agent