molecule-ai/hermes-agent
35
0
Fork 0
Code Issues 2 Pull Requests 4 Actions 14 Packages Projects Releases Wiki Activity

feat(tips): add cost-saving tips from April 30 tip-of-the-day (#17841)

Browse Source 
Seed the tips corpus with the knobs users can turn to reduce token
spend: hermes tools / hermes skills config to trim surface area,
/reasoning low|minimal to dial thinking depth down from the medium
default, and hermes models to route auxiliary tasks (vision, compression,
title gen, session_search) to cheaper backends while the main chat model
stays intact.

Requested by @micheltamanda under Teknium's tip-of-the-day tweet.
This commit is contained in:
Teknium 2026-04-30 02:30:36 -07:00 committed by GitHub
parent 0ad4f55aa8
commit 25caaa4a70
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
hermes_cli/tips.py
View File

@ -100,6 +100,9 @@ TIPS = [
"hermes gateway install sets up Hermes as a system service (systemd/launchd).",
"hermes memory setup lets you configure an external memory provider (Honcho, Mem0, etc.).",
"hermes webhook subscribe creates event-driven webhook routes with HMAC validation.",
"Save money: hermes tools disables unused tools, hermes skills config trims skills down.",
"/reasoning low or /reasoning minimal cuts thinking depth below the default (medium) — faster, cheaper responses.",
"hermes models routes vision, compression, and aux tasks to cheaper models — cuts background token cost 85%+ without downgrading your main chat model.",
# --- Configuration ---
"Set display.bell_on_complete: true in config.yaml to hear a bell when long tasks finish.",