1. Background Job Manager

   • Press key in emdx GUI → spawns background process
   • Each job gets a unique ID (maybe tied to doc ID)
   • Logs stream to files or database

2. TUI Log Viewer (new mode in emdx)

   • Press l for "logs" view
   • Shows active/recent jobs
   • Live tail of outputs
   • Filter by document, tag, status

def action_git_operation(self):
    job_id = f"git-{self.current_doc_id}-{timestamp}"
    
    subprocess.Popen([
        'bash', '-c', f'''
        exec > >(tee -a ~/.emdx/logs/{job_id}.log)
        exec 2>&1
        echo "=== Git Operation for: {doc['title']} ==="
        git checkout -b gameplan-{doc_id}
        git add .
        git commit -m "WIP: {doc['title']}"
        echo "=== Complete ==="
        '''
    ])
    
    status.update(f"Git operation started → Press 'l' to view logs")

def action_claude_execute(self):
    job = {
        'id': f"claude-{doc_id}-{timestamp}",
        'doc_id': doc_id,
        'type': 'claude-execution',
        'status': 'running',
        'log_file': f"~/.emdx/logs/claude-{doc_id}.log"
    }
    
    subprocess.Popen([
        'claude-code',
        '--file', temp_path,
        '--output-format', 'json',  # For structured parsing
        '--log-file', job['log_file']
    ])

Using something like Textual (which you already use), you could build:

┌─ EMDX Logs ──────────────────────────────────────────┐
│ Jobs (j/k to navigate, Enter to view)                │
├──────────────────────────────────────────────────────┤
│ ● [RUNNING] Claude: Implement Auth System      2m ago│
│ ✓ [DONE]    Git: Create feature branch        5m ago│
│ ✗ [FAILED]  Claude: Database Migration        10m ago│
│ ● [RUNNING] Monitor: API Health Check         Active │
└──────────────────────────────────────────────────────┘

┌─ Log Output ─────────────────────────────────────────┐
│ [13:45:02] Starting Claude execution...              │
│ [13:45:03] Analyzing document: Auth System Gameplan  │
│ [13:45:04] Found 5 implementation tasks              │
│ [13:45:05] Task 1/5: Create user model              │
│ [13:45:06] > Running: poetry add bcrypt             │
│ [13:45:08] > Success: bcrypt==4.0.1 installed       │
│ [13:45:09] > Creating: models/user.py               │
│ ▌                                                    │
└──────────────────────────────────────────────────────┘

1. Non-blocking - Can trigger multiple operations
2. Persistent logs - Review what happened later
3. Pattern matching - Search across all executions
4. Status tracking - See what's running/failed/completed
5. Less tmux clutter - Clean workspace

• stern - Multi-pod Kubernetes log tailing
• lnav - Advanced log file navigator
• multitail - Multiple log monitoring
• glogg - Fast log explorer

But building it into emdx means:

• Integrated with your documents

• Custom formatting for Claude/Git output

• Smart filtering by tags/projects

• Direct jump from log → document

• Press c in emdx GUI → Claude runs in background

• Press g → Git operations in background

• Press l → See all running/completed jobs

• Select a job → View live streaming logs

• Press d in log view → Jump back to source document

This creates a complete async workflow where you can trigger multiple operations and monitor them all from within emdx!