Getting Started
Aura Workshop ships in two shapes from one codebase: a desktop app for macOS, Windows, and Linux, and a headless daemon Docker image for servers. This page walks through installing either one, what happens on first launch, and how the application is laid out. It covers Aura Workshop 1.36.2.
Desktop app
Download the installer for your platform from the SecureInsights distribution channel
(control.secureinsights.ai). Filenames use the bare version — for example
Aura Workshop_1.36.2_aarch64.dmg, never a v prefix.
macOS
- Download
Aura Workshop_{version}_aarch64.dmg. - Open the DMG and drag the app into Applications.
- On first launch Gatekeeper may prompt. Shipped DMGs are Developer-ID signed and notarized, so a normal double-click opens cleanly.
- The Setup Wizard runs on first launch and downloads bundled runtimes
(Node, Python, uv, …) into
~/Library/Application Support/aura-workshop/bundled-deps/.
Requires macOS 12 or later. Apple Silicon only — there is no Intel build.
Windows
- Download
Aura Workshop_{version}_x64-setup.exe. - Run the NSIS installer. SmartScreen may warn; the installer is signed.
- First launch downloads bundled deps to
%AppData%\aura-workshop\bundled-deps\.
Requires Windows 10 or later (x64).
Linux
Choose AppImage (recommended, self-contained) or .deb (Debian / Ubuntu):
# AppImage
chmod +x "Aura Workshop_{version}_amd64.AppImage"
"./Aura Workshop_{version}_amd64.AppImage"
# .deb
sudo dpkg -i "Aura Workshop_{version}_amd64.deb"
Bundled deps land in ~/.local/share/aura-workshop/bundled-deps/.
The .deb declares its system dependencies (dpkg reports any missing):
libwebkit2gtk-4.1-0, libgtk-3-0,
libayatana-appindicator3-1, librsvg2-2, libsoup-3.0-0.
First launch & setup wizard
The first time you open Aura Workshop, a short setup wizard walks you through initial configuration. Every step is optional and revisitable later from Settings.
| Step | What it does |
|---|---|
| Language | Pick your interface language (10 locales are bundled). The choice persists across sessions. |
| License key | Enter a key to unlock paid tiers, or skip to continue. You can enter or upgrade a license at any time from Settings. |
| Provider + API key | Choose an AI provider and paste its API key. Add more providers later from the Models page. |
| Dependencies | Install the bundled runtimes (Node, Python, uv, and optional heavier deps) on demand. Skip and install later from Settings → General → Dependencies. |
| Composio sign-in (optional) | Connect cloud integrations through Composio if you plan to use the ~100-toolkit cloud OAuth mode. |
Behind the scenes
On first launch the app also initializes itself automatically:
- Creates a local SQLite database in WAL mode (63 base tables plus FTS5 search indexes) covering settings, conversations, tasks, teams, workflows, billing, credentials, and more.
- Detects whether Docker is installed. If not, it enables native mode so agent bash commands run directly on your OS.
- Installs the bundled skills library.
- Initializes credential encryption with an AES-256-GCM key held in the OS keychain (macOS Keychain, Windows Credential Manager, or the Linux secret service).
- Starts the embedded web server on port 18800 and auto-starts any listeners enabled in a prior session.
Where data lives
| Platform | Database path |
|---|---|
| macOS | ~/Library/Application Support/aura-workshop/aura-workshop.db |
| Windows | %AppData%\aura-workshop\aura-workshop.db |
| Linux | ~/.local/share/aura-workshop/aura-workshop.db |
| Daemon | /data/aura-workshop.db (via volume mount) |
File memory lives alongside the database under ~/.aura/memory/ (user scope).
Daemon (Docker)
The daemon runs the entire shared runtime — agent loop, tools, MCP, scheduler, listeners,
webhooks, workflows, and the SQLite database — inside a single container. It is published as
coolkoo/aura-workshop:daemon-latest on Docker Hub as a multi-arch manifest
spanning linux/amd64 and linux/arm64; Docker pulls the right variant
automatically.
Single container
docker run -d \
--name aura-daemon \
-p 18800:18800 \
-p 18790:18790 \
-v aura-data:/data \
-e AURA_WEB_TOKEN=$(uuidgen) \
-e AURA_API_KEY=sk-... \
-e AURA_MODEL=deepseek-chat \
-e AURA_BASE_URL=https://api.deepseek.com \
coolkoo/aura-workshop:daemon-latest
The daemon runs a single full mode. A --mode flag is accepted for
backward compatibility — worker is an alias, and any other value is coerced to
full.
Ports
| Port | Protocol | Purpose |
|---|---|---|
| 18800 | TCP | Web UI / viewer SPA + REST + SSE + /acp/* REST + Gateway /v1/* |
| 18790 | TCP | Inbound webhooks (separate from the main API server) |
| 18802 | UDP | ACP peer discovery broadcast |
The daemon SQLite lives at /data/aura-workshop.db (WAL mode) and file memory at
/root/.aura. Mount /data on a volume to persist across restarts.
Common environment variables
| Variable | Purpose |
|---|---|
AURA_WEB_TOKEN | Bearer token for /api/* and most /acp/* (env wins over the DB setting). |
AURA_API_KEY / AURA_MODEL / AURA_BASE_URL | Seed the default provider on first boot. |
AURA_VIEWER_MODE | Serve the viewer SPA instead of the full frontend. |
AURA_REMOTE_DEPLOYMENT | Strict auth — reject anonymous /api/* even when no token is set. |
AURA_HEARTBEAT_URL / AURA_DEPLOYMENT_ID | POST status every 30s plus a payload UUID. |
AURA_LICENSE_KEY | License key injected by the deployer. |
Reverse proxy + TLS
The daemon has no built-in TLS. Terminate HTTPS at Caddy (a Caddyfile ships in the
repo) or any reverse proxy:
aura.example.com {
reverse_proxy localhost:18800
}
For full daemon operations — docker compose, remote deployment over SSH, and the isolation model — see Deployment & Security.
Application layout
The desktop interface centers on a left sidebar — an always-visible icon rail for primary
navigation, a scrollable middle section for projects and tasks, and a main content area that
fills the rest of the screen. The daemon serves the same views in a browser at
http://your-daemon:18800/ (paste your AURA_WEB_TOKEN to sign in).
Left icon rail
The rail runs down the left edge and is always visible. Its top section holds fixed navigation icons for the primary views — the Dashboard (home screen with the composer, task history, and quick actions), the Agents / workspace (the main task surface for all AI interactions), and the automation views for Listeners, Webhooks, and Schedulers.
Projects & tasks
- Projects — a collapsible list of your projects. Each shows a name and a badge with its task count. Click + next to the Projects header to create one, or drag a task onto a project to associate it. See Projects for the full model.
- Tasks — a searchable, scrollable list of every task, each with a title and a colored status dot (completed, executing, failed, interrupted, waiting, planning). Click a task to open it in the main area.
Bottom section & behavior
Pinned to the bottom are Settings (the multi-tab settings panel), Report Bug, and Help / Docs. The sidebar is collapsible and resizable via the handle at its right edge; on narrow screens it folds into a hamburger menu that slides over the content.
Once you are oriented, head to the Workspace & Tasks guide to run your first task and learn the four task modes.
Where to next
- Workspace & Tasks — the composer and the four task modes (Execute, Plan, Goal, Auto).
- Projects — group tasks, files, and context with their own goals.
- Models & Providers — connect providers and set per-model parameters.
- Deployment & Security — the daemon image, remote deployment, and isolation.