Training & how to use Local AI
Get yourself or your team confident with the stack—Open WebUI, Obsidian, and local RAG. Personal use or team rollout; we cover the basics and best practices.
Personal and team training
Personal use
One person, one machine—your Mac or Linux box running Ollama and Open WebUI. Training focuses on: picking models, writing good prompts, using Obsidian as your knowledge base, and when to lean on RAG vs plain chat.
- First-time setup and model selection
- Chat best practices and prompt patterns
- Obsidian + Smart Connections + Copilot
- Adding docs to RAG and asking questions with sources
Team rollout
Several users sharing one Local AI host (e.g. Mac Mini or Linux server). Training covers: logging in to Open WebUI, roles and permissions, shared vs personal chats, and how to use the organisation’s RAG and Obsidian vaults.
- Access, roles, and safe use on a shared system
- Shared knowledge bases and document workflows
- Consistent prompts and use-case templates
- Who to contact for model or RAG updates
How to use Open WebUI
Open WebUI is your browser front-end for Local AI. You pick a model, type a message, and get a reply—all without leaving your machine. Training includes:
- Choosing the right model for the task (reasoning vs quick Q&A)
- Starting a chat, continuing threads, and when to start a new one
- Using RAG: selecting a knowledge base and asking questions that use your documents
- Understanding citations and how to check sources
- System prompts and persona-style instructions (if enabled by your admin)
How to use Obsidian with Local AI
Obsidian holds your notes and knowledge; plugins connect them to your local Llama model. Training covers:
- Opening the AI sidebar (Copilot or similar) and when to use it while writing
- Smart Connections: seeing related notes and asking the AI about links between them
- Text generation: summaries, expansions, and rewrites from your own wording
- Keeping sensitive content in your vault and never sending it to the cloud
Best practices we teach
- Prompt clearly: One clear question or instruction per message; add context when the AI needs it.
- Use RAG when it matters: For policy, docs, or project knowledge, point the chat at the right knowledge base so answers are grounded in your data.
- Check citations: When the AI quotes a source, skim the original to confirm it’s correct.
- Respect roles: On shared systems, stick to your permissions and don’t change model or RAG settings unless you’re allowed.
Training we offer
We run sessions tailored to your setup: personal (one user) or team (several users on one Local AI host). Typically a mix of live walkthrough and Q&A. We can also leave short written guides or short videos for your team to reuse.
Request training