Local AI Agents
Bounded agents you can trust on your own machine.
Self-paced · Evidence-led · Built in the open
Practical AI learning for responsible operators.
Learn the systems behind useful AI: bounded agents, protected data, verifiable evidence and accountable decisions. 4 free, self-paced courses — open now. No accreditation claims, no hype — verified completion certificates only.
- Evidence-ledShow what happened.
- Safe by designBound tools and data.
- Operator firstHumans stay accountable.
Learning system 01Local AI AgentsTools · Memory · Evidence · Rollback
Three focused paths
Choose your learning route
AI Foundations
What modern AI can do, where it fails, and how to work with it safely.
AI-Powered Project Management
AI in the project room — with governance and human accountability intact.
Local AI Agents: Advanced Systems
Multi-agent systems, injection defence and evaluation — still on your own hardware.
Monday · Wednesday · Friday
Articles worth your time
A web page is not permission for your AI agent
An agent can use a page, document or message as evidence without treating instructions inside it as authority. Design that boundary before connecting powerful tools.
Automate disk cleanup only after you can explain the growth
A full disk invites a hurried deletion. A safer home-server workflow identifies the owner, growth rate, reclaimable data, retention need and recovery path before cleanup becomes automatic.
Local AI stays private only if the whole workflow does
Running a model on your own machine can remove one remote data path, but prompts, tools, logs, saved outputs, backups and sharing still decide where information travels.
What Learning Harbour is — and isn't
- Live now
- 4 free, self-paced courses, open to sign in and start today.
- Certificates
- State verified platform completion only — no accreditation, no identity verification, no employment or income claims.
- Not claimed
- No learner outcomes, no testimonials, no reviews.
- Still in development
- Remaining course content, written and reviewed one module at a time.