A hands-on introduction to self-improving AI agents and live financial data. Build an agent that queries Indonesian stock market data and gets better the more you use it.
Docs and downloads: hermes-agent.nousresearch.com and github.com/NousResearch/hermes-agent.
Hermes Agent is an open-source AI agent from Nous Research that learns from experience. It writes its own skills, revises them when it finds a better approach, and keeps a persistent record of your preferences. Connected to the Sectors API, it becomes a financial assistant that improves at answering your questions over time.
By the end, you'll have a working Hermes instance connected to a messaging platform and live market data, plus a scheduled financial workflow running on its own.
From foundations to a fully running, scheduled financial agent.
| # | Module | Focus |
|---|---|---|
| 01 | What is Hermes? | Agents vs chatbots, honest expectations |
| 02 | Hermes Architecture | Self-improvement loop, memory files, security |
| 03 | Setup & Installation | Install, SOUL.md, Nous Portal, Telegram & Discord |
| 04 | Sectors MCP | API key, MCP connection, first query |
| 05 | Memory & Skills | SOUL.md, MEMORY.md, USER.md, watching a skill get created |
| 06 | Automation & Cron | Cron workflows, self-contained prompts, VPS |
| 07 | Multi-Agent Orchestration | Kanban board, delegation, the working demo pipeline |
| 08 | Level Up | SOUL.md tuning, skill curation, OpenClaw migration |
By the end of this workshop you will be able to:
Explain a chatbot vs a self-improving agent to a non-technical colleague.
Install Hermes and connect it to a model provider and a messaging channel.
Connect Sectors and retrieve live Indonesian stock data through natural language.
Understand how Hermes writes and refines skills, and how to curate them.
Schedule recurring financial workflows, briefings, snapshots, alerts, that run on their own.