Videos are part of the 10-part Hermes Agent Masterclass
Curated by: Tonbi's AI Garage (11 videos)
Module 5 of the Hermes Agent Masterclass: every skill and every task runs on a model — here's how to configure them, cut costs, run local, and serve your subscription to any tool. Sign up for my FREE weekly newsletter, where I spill my unfiltered thoughts on the latest AI news, cool research, and projects I'm building: https://www.onchainaigarage.com/ 🐦 Follow Tonbi on X for real-time AI x blockchain updates! https://x.com/tonbistudio Module 5 maps the full model and provider landscape in Hermes Agent — 28 first-class plugins, API keys vs OAuth subscriptions, and how to layer cost tiers using fallback providers, per-task auxiliary models (vision, title generation, session search, and more), and credential pools that rotate API keys automatically to survive rate limits. Then I get practical: running a 9B model locally on an RTX 3060 with llama.cpp (quantized KV cache, flash attention, Jinja template for tool calling — the flags that actually fit a big model into 12GB), a simpler Ollama path with the one context-length env variable you must set or your agent will break, and the new Hermes Proxy that turns your existing OAuth subscription into a local OpenAI-compatible endpoint any other tool can hit — Cline, Continue, Cursor, aider, or your own scripts. Timestamps: 0:00 - Why models matter: every skill and task runs on one 1:12 - Provider landscape: API keys vs OAuth explained 5:21 - Cost tiering: fallback models and auxiliary model slots 10:16 - Credential pools and rate-limit resilience 12:25 - Switching models mid-session and setting reasoning effort 13:38 - Local models: llama.cpp on an RTX 3060 (9B in 12GB) 17:32 - Ollama setup and the context length caveat 20:04 - Hermes Proxy: serve your subscription as any tool's backend What are you running as your local model in Hermes — Ollama, llama.cpp, something else? And what auxiliary slot did you customize first? Drop it in the comments. Module 6 is on Tools and MCP, so subscribe and hit the bell so you don't miss it! 🦐✨ #HermesAgent #LocalLLM #llamacpp #Ollama #AIAgents #AITools #LocalAI #OpenSourceAI #Masterclass