Getting Started

Install and run Omnideck in under five minutes.

Omnideck runs inside a container managed by the omnideck CLI. You need a container engine (Podman or Docker) and either a cloud LLM API key or a local Ollama install.

1. Install the CLI — Homebrew (recommended) or binary download
2. Install Omnideck — the setup wizard handles the rest
3. Open the tool — connect your LLM and start working

Prerequisites

• Container engine: Podman 4.0+ (recommended) or Docker 20.10+. Other container engines such as Colima are not supported at this time. We recommend Podman for its open-source nature and smaller resource footprint.
• LLM provider: an API key for OpenAI, Anthropic, OpenRouter, or any OpenAI-compatible endpoint — or Ollama for local models
• RAM: 4 GB minimum (8 GB recommended for local models)

Step 1: Install the CLI

The omnideck CLI wraps your container engine with a guided installer and simple management commands.

Option A: Homebrew (macOS and Linux)

The recommended way to install on macOS or Linux:

brew install omnideck-dev/tap/omnideck

To upgrade later:

brew upgrade omnideck

Option B: Download a binary

Grab the binary for your OS from the releases page:

Extract the archive and move the binary to a directory on your PATH. On Linux or macOS:

tar -xzf omnideck-linux-amd64.tar.gz
chmod +x omnideck
sudo mv omnideck /usr/local/bin/

Verify the install

omnideck --version

Step 2: Install Omnideck

omnideck install

The install wizard:

1. Detects your container engine (Docker or Podman)
2. Checks whether Ollama is reachable on the host
3. Suggests container memory limits sized for your system
4. Pulls the container image and starts the container

When the wizard finishes, open http://localhost:2337 in your browser.

Step 3: Open the tool

A setup wizard runs in the UI the first time you open it. It guides you through:

1. Adding an LLM provider — cloud (OpenAI, Anthropic, OpenRouter, or any OpenAI-compatible endpoint) or local Ollama
2. Picking your main model — used for chat, context compaction, and conversation titles
3. Picking an optional vision model — for tasks that involve image input

Cloud providers list available models automatically. Ollama lists whatever you've already pulled.

That's it — you're running.

Managing Omnideck

Everything is driven through the CLI:

omnideck status      # container state, data dirs, Ollama reachability, and web UI port
omnideck stop        # gracefully stop the container
omnideck start       # start it back up
omnideck restart     # stop then start
omnideck update      # pull the latest image and recreate the container
omnideck logs -f     # tail container logs
omnideck doctor      # run health checks and print a pass/warn/fail report
omnideck uninstall   # remove the container (optionally back up and delete data)

Your data lives in ~/Omnideck and survives restarts and upgrades. Conversations, memory, agent profiles, goals, and generated files are all preserved when you update.

Next steps

• CLI Reference — full command and flag reference
• Local Models — run Omnideck fully offline with Ollama
• Integrations — connect Gmail, Calendar, and Drive
• Agents — create and customize agent profiles
• Autonomous Tasks — schedule background goals