dubnium0/hermes-profile-insights

Hermes Insights is a Python script that reads the local data files created by Hermes, an AI assistant platform, and produces terminal reports showing how each profile is being used. If you run Hermes with multiple profiles, such as one for different projects or clients, this tool lets you see what each profile is doing at a glance without opening the databases individually. The reports cover session and message counts, which AI models were used and how many tokens each consumed, which tools and skills were invoked, how active each profile has been over time including peak hours and active day streaks, and which individual sessions stand out as longest or most intensive. A cost report mode shows estimated spending over a configurable number of recent days. Cost figures come from two sources that the tool keeps separate. Stored cost uses whatever cost value Hermes already recorded in the database. Computed cost uses built-in pricing rates for OpenAI and Gemini models to calculate from token counts. You can add a JSON file with custom prices to override the built-in rates or cover models that are not included. The tool opens databases in read-only mode, so it cannot change anything in your Hermes installation. You run it either interactively, where it lets you pick a profile from a menu, or by passing the profile name and a report mode directly on the command line. The design notes in the README emphasize keeping everything local and transparent: no data leaves your machine, the source of each cost number is labeled clearly, and the tool avoids guessing or scoring behavior in ways that are not grounded in the raw numbers. No license is specified in the repository.