What is it?

Local AI Workspace is an installer kit (install.sh script + automatic configuration) that gets a working AI environment running on your own computer, with zero dependency on any cloud service.

What problem does it solve?

Setting up local AI by hand means picking models, installing Ollama, configuring Docker, standing up AnythingLLM, and creating workspaces one by one: easily a lost afternoon. This kit automates the whole process.

What’s included

install.sh: installs Ollama, downloads the recommended models, and spins up AnythingLLM in Docker.
3 pre-configured workspaces, created via the AnythingLLM REST API: Chat (direct conversation with the local model), RAG (upload your own documents and ask questions about them; nothing leaves your machine) and Agent (task-oriented workspace with tools).
Pre-tuned recommended models: gemma3:4b (general chat), qwen2.5-coder:3b (code/agent), nomic-embed-text (RAG embeddings).
.env template to optionally add external providers (OpenAI/Claude) later.
Post-install verification checklist.
PDF guide with a model comparison table and FAQ.

Why 100% local

None of your data ever leaves your machine. No mandatory paid API keys, no third-party usage limits, no monthly subscription.

Requirements

Docker installed, ~6 GB free for the default models, Linux/macOS/WSL recommended.

Not included (v1.0)

Multi-user support/authentication (built for personal/LAN use).
Cloud hosting: by design, this is 100% local.

Software language

Available in English and Spanish (bilingual ZIP with separate ES / EN variants).

JAH.STORE transparency

Scripts are syntax-verified, but the full clean-install experience depends on each user’s environment; any issue found is documented and fixed in the next version.

RAG

Local AI Workspace – Your 100% Local AI, Running in Under 15 Minutes (Ollama + AnythingLLM)