Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/TrinaxCode/TrinaxAI/llms.txt

Use this file to discover all available pages before exploring further.

TrinaxAI runs entirely on your Windows machine — no cloud, no subscriptions, no data leaving your network. The PowerShell installer (install.ps1) auto-detects your RAM, writes .env, installs Python, Node.js, and Git via winget if they are missing, installs Ollama using winget with a silent-installer fallback, builds the PWA frontend, configures Windows Firewall rules for LAN access, and optionally registers TrinaxAI in the Windows Startup folder. Services run as supervised background processes — no separate service manager installation is required. When the install completes you’ll have:
  • Ollama running at http://localhost:11434
  • RAG API (FastAPI) at https://localhost:3333
  • PWA chat interface at https://localhost:3334
PowerShell’s default execution policy blocks unsigned scripts. Run install.ps1 with the -ExecutionPolicy Bypass flag — this applies only to the current PowerShell session and does not permanently change your system policy.
powershell -ExecutionPolicy Bypass -File .\install.ps1

​
Install steps

1

Open PowerShell

Press Win + X and choose Windows PowerShell or Terminal. You do not need to run as Administrator for a basic install — the installer will prompt if elevated rights are needed for firewall rules.
2

Clone the repository

git clone https://github.com/TrinaxCode/TrinaxAI.git $env:USERPROFILE\trinaxai
cd $env:USERPROFILE\trinaxai
If Git is not installed yet, the installer will install it via winget automatically. You can also download it from git-scm.com and reopen your terminal.
3

Run the installer

powershell -ExecutionPolicy Bypass -File .\install.ps1
The installer will:
  1. Detect your RAM and select a hardware profile
  2. Write .env with your auto-detected settings
  3. Install Python, Git, Node.js, and Ollama via winget (with silent-installer fallback for Ollama)
  4. Create a Python .venv and install all packages
  5. Build the PWA frontend with npm
  6. Generate and trust a local HTTPS certificate
  7. Add Windows Firewall rules for ports 3333 and 3334 on Private networks
  8. Prompt to download Ollama models
  9. Prompt to enable Windows startup autostart
  10. Optionally start TrinaxAI now
4

Open the PWA

https://localhost:3334
Accept the local certificate warning on first visit. If you chose to trust it automatically during install, Chrome and Edge will open without a warning.

​
Installer flags

# Fully automatic, no prompts
powershell -ExecutionPolicy Bypass -File .\install.ps1 -NonInteractive

# Skip model downloads
powershell -ExecutionPolicy Bypass -File .\install.ps1 -NoModels

# Skip vision model only
powershell -ExecutionPolicy Bypass -File .\install.ps1 -NoVision

# Do not enable Windows startup
powershell -ExecutionPolicy Bypass -File .\install.ps1 -NoAutostart

# Do not start services at the end
powershell -ExecutionPolicy Bypass -File .\install.ps1 -NoStart

# Enable LAN system-control endpoints
powershell -ExecutionPolicy Bypass -File .\install.ps1 -LanSystem

# Force a hardware profile
powershell -ExecutionPolicy Bypass -File .\install.ps1 -Profile 16gb
Flag reference
FlagDescription
-InteractiveGuided install; prompts for optional choices (default)
-NonInteractiveFully automatic — no prompts. Useful for CI/scripts
-NoModelsSkip downloading all Ollama models (including vision)
-NoVisionSkip vision model download only
-NoAutostartDo not register TrinaxAI in Windows Startup
-NoStartDo not start TrinaxAI at the end of the install
-ProfileOverride auto-detected profile: 8gb, 16gb, max, or ultra
-LanSystemEnable LAN system-control endpoints and generate an admin token
Profile reference
ProfileRAM targetModels used
8gb≤ 8 GBllama3.2:1b, qwen2.5-coder:1.5b, nomic-embed-text
16gb9–19 GBllama3.2:3b, qwen2.5-coder:3b, bge-m3
max20–31 GBqwen2.5-coder:7b, larger context window
ultra≥ 32 GBqwen2.5-coder:14b, qwen2.5vl:7b

​
Installing dependencies manually

If winget is unavailable, install each tool manually and reopen your terminal before running the installer: 
winget install --id Python.Python.3.12 --silent
winget install --id Git.Git --silent
winget install --id OpenJS.NodeJS.LTS --silent
winget install --id Ollama.Ollama --silent
After installing, verify everything is on your PATH: 
python --version
git --version
node --version
npm --version
ollama --version

​
Service management

Start, stop, and check status using the Python service manager: 
# Start all services
.\.venv\Scripts\python.exe service_manager.py start --base-dir "$PWD"

# Shut down AI (RAG + Ollama); PWA stays available
.\.venv\Scripts\python.exe service_manager.py stop-ai --base-dir "$PWD"

# Shut down everything
.\.venv\Scripts\python.exe service_manager.py stop-all --base-dir "$PWD"

# Check status
.\.venv\Scripts\python.exe service_manager.py status --base-dir "$PWD"

​
Autostart on Windows boot

The installer optionally registers a TrinaxAI.vbs launcher in the Windows Startup folder so no console window stays visible on login. Manage it with: 
# Enable autostart
.\.venv\Scripts\python.exe service_manager.py enable-autostart --base-dir "$PWD"

# Disable autostart
.\.venv\Scripts\python.exe service_manager.py disable-autostart --base-dir "$PWD"

# Browse the Startup folder manually
explorer "$env:APPDATA\Microsoft\Windows\Start Menu\Programs\Startup"
The supervisor keeps the PWA available at all times. If you shut down the AI from the PWA settings or with stop-ai, the next Windows startup will not restart Ollama/RAG — it respects the state you left it in.

​
Windows Firewall and LAN access

The installer adds inbound rules for TCP ports 3333 and 3334 on Private networks when run with Administrator privileges. If the rules weren’t applied (non-admin install), add them manually: 
# Run PowerShell as Administrator, then:
New-NetFirewallRule -DisplayName "TrinaxAI RAG API" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 3333 -Profile Private
New-NetFirewallRule -DisplayName "TrinaxAI PWA"     -Direction Inbound -Action Allow -Protocol TCP -LocalPort 3334 -Profile Private
Accessing from a phone on the same Wi-Fi: open https://[YOUR-LAN-IP]:3334 in your phone’s browser. Find your Windows LAN IP: 
Get-NetIPAddress -AddressFamily IPv4 |
  Where-Object { $_.IPAddress -match "^(192\.168|10\.|172\.(1[6-9]|2[0-9]|3[0-1]))" } |
  Select-Object -First 1 IPAddress
Do not set Firewall rules for Public networks. Do not expose ports 3333, 3334, or 11434 to the internet. For remote access, use a VPN such as Tailscale or WireGuard.

​
Update

Use the native Windows updater. It asks whether to back up first, pull the latest code, update models, change autostart, and restart services. Python and npm dependencies always update automatically: 
cd $env:USERPROFILE\trinaxai
powershell -ExecutionPolicy Bypass -File .\update.ps1

​
Uninstall

Use the native Windows uninstaller. It stops services, disables autostart, and asks which generated/runtime files to remove. Your source code is always kept: 
cd $env:USERPROFILE\trinaxai
powershell -ExecutionPolicy Bypass -File .\uninstall.ps1
RAG index data (storage\) and uploaded files (local_sources\) are kept by default. The uninstaller will ask whether to remove them during the guided flow.

​
Port reference

PortServiceNotes
3333RAG API (FastAPI)Backend — binds to 0.0.0.0
3334PWA (chat interface)Open https://localhost:3334 in your browser
11434OllamaLocal model engine
LAN system control is disabled by default. Sensitive endpoints require either a localhost caller or a valid TRINAXAI_ADMIN_TOKEN. Enable it during install with -LanSystem, or set TRINAXAI_ALLOW_LAN_SYSTEM=1 and TRINAXAI_ADMIN_TOKEN in .env manually.

​
Note on WSL2

You can run TrinaxAI inside WSL2 using the Linux guide, but the most direct path on Windows is PowerShell + install.ps1. If you use WSL2, keep in mind that networking, firewall rules, and file-system access behave differently between the Windows and Linux layers.

Build docs developers (and LLMs) love

Get started for freeTalk to us