The Continuity Layer
Run your own AI · The learning path

From zero to your own AI, in four levels

thecontinuitylayer.com/open/learn

Checked 6 November 2026. US dollars. No affiliate links.

No coding needed for levels one and two. Each step ends with something you can do, not just something you know. Free, no sign-up.

Level 1: Understand your options

Twenty minutes. You will know what you are choosing between, and what each choice costs you.

  1. 1.1
    Open weight, open source and the rest

    Read any model's license and know what you may do with it

  2. 1.2
    What "uncensored" AI really means

    Tell a maker's model from a community retune, and know what changes

  3. 1.3
    Local or hosted: the honest trade-offs

    Decide whether running your own is worth it for you

  4. 1.4
    Are Chinese models safe?

    Know where your words go with Qwen, DeepSeek, GLM and Kimi

Level 2: Get started

One evening. You will have a model running on your own computer and a chat window on top of it.

  1. 2.1
    Run your first local model

    A private model answering you, offline, in about fifteen minutes

  2. 2.2
    Model sizes and quantization

    Look at any download and know whether it will fit

  3. 2.3
    Pick a model with the finder

    The right model for the memory you have

  4. 2.4
    Pick your app

    Ollama, LM Studio, Jan or Open WebUI, chosen for how you work

  5. 2.5
    Why is it slow or broken?

    Fix the five problems almost everyone hits

  6. 2.6
    The hardware guide

    Know what your computer can run, and what to upgrade if you want more

Level 3: Take control

Make the model behave your way and work on your own files.

  1. 3.1
    Switch models, keep your context

    Your instructions and notes in one file that any model can read

  2. 3.2
    System prompts and custom models

    Save a model with your own standing instructions

  3. 3.3
    Chat with your own documents

    Ask questions of a folder of PDFs without uploading them anywhere

  4. 3.4
    Updates and new versions

    Swap in a newer model without losing your setup

Level 4: Build your own AI environment

From one person's chat window to a setup the household, or the business, can use.

  1. 4.1
    Give your AI a memory it keeps

    A record your AI reads first and writes to last, whichever model you run

  2. 4.2
    When to rent a GPU instead

    Run the big models by the hour without buying hardware

  3. 4.3
    One key for every model

    Mix local models with hosted ones behind one API

  4. 4.4
    A local AI server for the house

    Open WebUI with accounts, on one machine everyone can reach

  5. 4.5
    Agents on local models

    Tools and automations that never leave your network