feat(integrations): add n8n automation integration

src-tauri/src/core/system/commands.rs

@@ -3590,6 +3590,27 @@ pub fn open_agent_terminal(
     Ok(())
 }
 
+/// Open a URL in the user's default browser (cross-platform).
+fn open_web_url(url: &str) -> Result<(), String> {
+    use std::process::Command;
+
+    let result = if cfg!(target_os = "windows") {
+        Command::new("cmd")
+            .args(["/C", "start", "", url])
+            .status()
+    } else if cfg!(target_os = "macos") {
+        Command::new("open").arg(url).status()
+    } else {
+        Command::new("xdg-open").arg(url).status()
+    };
+
+    match result {
+        Ok(status) if status.success() => Ok(()),
+        Ok(_) => Err(format!("Failed to open {}", url)),
+        Err(e) => Err(format!("Failed to open {}: {}", url, e)),
+    }
+}
+
 /// Launch a GUI editor for the "IDEs & Editors" integrations (VS Code,
 /// JetBrains, Xcode).
 ///
@@ -3641,6 +3662,23 @@ pub fn launch_editor(editor_id: String) -> Result<(), String> {
         // Xcode is macOS-only and ships no general-purpose launcher binary
         // (`xed` needs a file argument), so we open the app directly.
         "xcode" => (&[], Some("Xcode")),
+        // Self-hosted web UIs — open the default local URL in the browser.
+        "onyx" => {
+            open_web_url("http://localhost:3000")?;
+            log::info!("Opened Onyx at http://localhost:3000");
+            return Ok(());
+        }
+        "n8n" => {
+            for cli in &["n8n"] {
+                if try_spawn(cli, &["start"]) {
+                    log::info!("Launched n8n via '{} start'", cli);
+                    return Ok(());
+                }
+            }
+            open_web_url("http://localhost:5678")?;
+            log::info!("Opened n8n UI at http://localhost:5678");
+            return Ok(());
+        }
         other => return Err(format!("Unknown editor: {}", other)),
     };
 

web-app/src/constants/integrations.ts

@@ -1,4 +1,18 @@
-export type IntegrationKind = 'coding' | 'assistant' | 'editor'
+export type IntegrationKind =
+  | 'coding'
+  | 'assistant'
+  | 'editor'
+  | 'chat'
+  | 'automation'
+  | 'notebook'
+
+/** Integrations that use copy-settings + launch + manual steps (no config file). */
+export const MANUAL_SETUP_KINDS: IntegrationKind[] = [
+  'editor',
+  'chat',
+  'automation',
+  'notebook',
+]
 
 /**
  * Ordered manual-setup instructions shown inside an editor card. These mirror
@@ -121,20 +135,6 @@ export const INTEGRATION_AGENTS: IntegrationAgent[] = [
     requiresModel: true,
     endpointWithPrefix: true,
   },
-  {
-    id: 'zed',
-    name: 'Zed',
-    description: 'High-performance code editor with a built-in AI agent.',
-    kind: 'coding',
-    detectBin: 'zed',
-    docsUrl: 'https://zed.dev/docs/ai/llm-providers',
-    installable: true,
-    configurable: true,
-    requiresModel: true,
-    // Zed's native Atomic Chat provider expects the OpenAI-compatible base URL
-    // with the `/v1` prefix (matches its built-in default).
-    endpointWithPrefix: true,
-  },
   {
     id: 'mimo',
     name: 'MiMo Code',
@@ -285,4 +285,48 @@ export const INTEGRATION_AGENTS: IntegrationAgent[] = [
       ],
     },
   },
+  {
+    id: 'onyx',
+    name: 'Onyx',
+    description:
+      'Self-hostable chat UI with RAG, agents, connectors, and deep research.',
+    kind: 'chat',
+    detectBin: 'docker',
+    docsUrl: 'https://docs.onyx.app/deployment/quickstart',
+    installable: false,
+    configurable: false,
+    requiresModel: false,
+    endpointWithPrefix: true,
+    editor: {
+      launchId: 'onyx',
+      steps: [
+        'Deploy Onyx (see Docs) and sign in to your instance.',
+        'During setup, add an OpenAI-compatible LLM provider.',
+        'Paste the copied Base URL and API key (use host.docker.internal instead of 127.0.0.1 if Onyx runs in Docker).',
+        'Select your Atomic Chat model and send a test message.',
+      ],
+    },
+  },
+  {
+    id: 'n8n',
+    name: 'n8n',
+    description:
+      'Workflow automation with AI nodes powered by your local models.',
+    kind: 'automation',
+    detectBin: 'n8n',
+    docsUrl: 'https://docs.n8n.io/integrations/builtin/credentials/ollama/',
+    installable: false,
+    configurable: false,
+    requiresModel: false,
+    endpointWithPrefix: false,
+    editor: {
+      launchId: 'n8n',
+      steps: [
+        'Open n8n → Credentials → Add credential → Ollama (or OpenAI with a custom base URL).',
+        'Set Base URL to the copied value (use host.docker.internal if n8n runs in Docker).',
+        'Save and test the connection.',
+        'Add an Ollama / OpenAI chat node to a workflow and pick your Atomic Chat model.',
+      ],
+    },
+  },
 ]

web-app/src/locales/en/launch.json

@@ -9,6 +9,12 @@
   "assistantsDesc": "AI assistants that help with everyday tasks.",
   "editors": "Editors",
   "editorsDesc": "Code editors that let you chat and code with your local models.",
+  "chatRag": "Chat & RAG",
+  "chatRagDesc": "Chat interfaces and retrieval-augmented generation platforms.",
+  "automation": "Automation",
+  "automationDesc": "Workflow automation platforms with AI integration.",
+  "notebooks": "Notebooks",
+  "notebooksDesc": "Interactive notebooks with AI chat and code completion.",
   "copySettings": "Copy settings",
   "manualSetup": "Manual setup",
   "serverNotRunningHint": "Start the local server above first so {{name}} can reach your models.",
@@ -18,8 +24,7 @@
   "setCustomPath": "Set binary path",
   "customPathSet": "Binary path set",
   "savePath": "Save",
-  "customPathPlaceholder": "/path/to/binary",
-  "customPathPlaceholderWindows": "C:\\path\\to\\binary.exe",
+  "customPathPlaceholder": "/path/to/binary or C:\\path\\to\\binary.exe",
   "customPathHint": "Point at the agent's executable if it's installed in a non-standard location that auto-detection misses. This only affects the Installed status here.",
   "install": "Install",
   "installing": "Installing...",
@@ -44,10 +49,8 @@
     "installSuccess": "{{name}} installed",
     "installSuccessDesc": "{{name}} was installed successfully.",
     "installFailed": "Failed to install {{name}}",
-    "installFailedNetworkDesc": "The installer couldn't reach the network. If you're behind a region block, set an HTTP/HTTPS proxy in Settings → HTTPS Proxy and try again. Note: SOCKS proxies often don't work for npm/PowerShell-based installers.",
     "configured": "{{name}} configured",
     "configuredDesc": "{{name}} now points at your local server — opening a terminal.",
-    "configuredDescEditor": "{{name}} now points at your local server — opening the editor.",
     "configureFailed": "Failed to configure {{name}}",
     "terminalFailed": "Couldn't open a terminal for {{name}}",
     "settingsCopied": "Connection settings copied",