Merge remote Gitea history

.gitignore

@@ -0,0 +1,61 @@
+# Models & Large Files
+.ollama/
+models/
+*.safetensors
+*.gguf
+*.bin
+*.pth
+
+# Workspace
+workspace/
+uploads/
+backups/
+
+# Python
+__pycache__/
+*.pyc
+*.pyo
+*.egg-info/
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+.venv/
+venv/
+ENV/
+
+# IDE
+.vscode/
+.idea/
+*.swp
+*.swo
+*~
+.DS_Store
+
+# Environment
+.env
+.env.local
+.env.*.local
+
+# Logs
+logs/
+*.log
+
+# Temporary
+*.tmp
+*.cache
+.pytest_cache/
+.coverage
+
+# OS
+.DS_Store
+Thumbs.db

Makefile

@@ -0,0 +1,89 @@
+.PHONY: help install update monitor health-check backup clean reinstall logs
+
+help:
+	@echo "LLM Server - Comandos disponibles:"
+	@echo ""
+	@echo "  make install          - Instalar todo desde cero (requiere sudo)"
+	@echo "  make update           - Actualizar código y dependencias"
+	@echo "  make monitor          - Ver monitoreo en tiempo real"
+	@echo "  make health-check     - Verificar estado del sistema"
+	@echo "  make backup           - Hacer backup de configuración"
+	@echo "  make logs             - Ver logs de servicios"
+	@echo "  make status           - Ver estado de servicios"
+	@echo "  make start            - Iniciar servicios"
+	@echo "  make stop             - Detener servicios"
+	@echo "  make restart          - Reiniciar servicios"
+	@echo "  make clean            - Limpiar archivos temporales"
+	@echo "  make reinstall        - Reinstalar (redownload modelos)"
+
+install:
+	@echo "Instalando LLM Server..."
+	sudo bash scripts/install.sh
+
+install-quick:
+	@echo "Instalación rápida (sin modelos)..."
+	sudo bash scripts/install.sh --quick
+
+install-gpu-only:
+	@echo "Instalando solo GPU drivers..."
+	sudo bash scripts/install.sh --gpu-only
+
+update:
+	bash scripts/update.sh
+
+monitor:
+	bash scripts/monitor.sh
+
+health-check:
+	bash scripts/health-check.sh
+
+backup:
+	bash scripts/backup.sh
+
+logs:
+	@echo "Logs de Ollama:"
+	sudo journalctl -u ollama -f
+
+logs-api:
+	@echo "Logs de API:"
+	sudo journalctl -u llm-api -f
+
+status:
+	@echo "Ollama:"
+	@sudo systemctl status ollama --no-pager
+	@echo ""
+	@echo "LLM API:"
+	@sudo systemctl status llm-api --no-pager
+
+start:
+	sudo systemctl start ollama llm-api
+	@echo "✅ Servicios iniciados"
+
+stop:
+	sudo systemctl stop ollama llm-api
+	@echo "✅ Servicios detenidos"
+
+restart:
+	sudo systemctl restart ollama llm-api
+	@echo "✅ Servicios reiniciados"
+
+clean:
+	rm -rf __pycache__ .pytest_cache .venv venv
+	find . -type f -name "*.pyc" -delete
+	@echo "✅ Limpiado"
+
+reinstall:
+	@echo "Reinstalando LLM Server..."
+	sudo bash scripts/install.sh --gpu-only
+	make update
+	@echo "✅ Reinstalación completada"
+
+info:
+	@echo "========== INFORMACIÓN DEL SERVIDOR =========="
+	@echo "Hostname: $$(hostname)"
+	@echo "IP: $$(hostname -I)"
+	@echo "SO: $$(lsb_release -d | cut -f2)"
+	@echo "CPU: $$(lscpu | grep 'Model name' | cut -d: -f2)"
+	@echo "RAM: $$(free -h | awk 'NR==2 {print $$2}')"
+	@echo "GPU: $$(nvidia-smi --query-gpu=name --format=csv,noheader 2>/dev/null || echo 'No disponible')"
+	@echo "=============================================="

README.md

@@ -1,5 +1,92 @@
+# LLM Server - Deployment Automatizado
+
+Sistema completo de IA agentica con LLMs locales (Phi + DeepSeek 33B) en Ubuntu 22.04.
+
+## Requisitos Mínimos
+
+- **CPU**: Intel i3-10105 (4 cores)
+- **RAM**: 16GB DDR4
+- **GPU**: NVIDIA RTX 3090 (24GB VRAM)
+- **PSU**: 850W+
+- **SO**: Ubuntu 22.04 LTS
+
+## Solución para tool calling con rutas de Windows
+
+Cuando el LLM intenta ejecutar una herramienta, la respuesta no debe mostrarse como texto plano. Debe ser manejada como un `tool_call` o `function_call` y ejecutada localmente.
+
+### Regla importante
+
+Añade esta instrucción al `System Prompt`:
+
+```text
+When outputting Windows file paths in JSON arguments, you must strictly escape all backslashes (for example: C:\\Users\\name\\file.txt).
+```
+
+### Patrón recomendado
+
+1. El LLM devuelve una llamada como JSON.
+2. Tu cliente detecta `tool_calls` o `function_call`.
+3. Tu código ejecuta la función localmente.
+4. Se envía de vuelta al LLM el resultado con rol `tool` o `function`.
+
+### Ejemplo de implementación
+
+El proyecto incluye un ejemplo funcional en `tool_call_demo.py` que:
+
+- simula un `tool_call` del modelo,
+- valida que la ruta JSON esté escapada correctamente,
+- lee el archivo físico en disco,
+- devuelve el contenido al modelo para continuar.
+
+```bash
+python tool_call_demo.py
+```
+
+### Ejemplo de payload válido
+
+```json
+{
+  "tool_calls": [
+    {
+      "id": "call_read_001",
+      "type": "function",
+      "function": {
+        "name": "read_file",
+        "arguments": "{\"path\": \"c:\\\\Workspace\\\\llm-server-setup\\\\README.md\"}"
+      }
+    }
+  ]
+}
+```
+
+> Si el JSON tiene barras invertidas sin escape, el parser falla y el modelo parece "no responder". Usar rutas con `\\` en JSON es obligatorio en Windows.
+
+## Instalación Rápida
+
+```bash
+# 1. Clonar repositorio
+git clone https://github.com/tuuser/llm-server-setup.git
+cd llm-server-setup
+
+# 2. Instalar OpenClaw sin iniciar el asistente
+bash install_openclaw.sh
+
+# 3. Ejecutar el asistente inicial cuando quieras
+openclaw onboard --install-daemon
+
+# También puedes instalar y abrir el asistente en el mismo paso
+bash install_openclaw.sh --onboard
+```
+
+El script está pensado para Ubuntu 22.04/24.04 y WSL2. Usa el instalador oficial
+de OpenClaw, que instala Node.js si es necesario. Revisa las instrucciones
+oficiales si vas a usar otra distribución o un entorno restringido.
+
+## Instalación del servidor LLM
+
+```bash
 nano setup_llm_server.sh
 
 chmod +x setup_llm_server.sh
 sudo ./setup_llm_server.sh
-
+```

agent.py

@@ -0,0 +1,316 @@
+"""
+Code Agent - Handles file operations, code execution, and testing
+"""
+
+import os
+import json
+import requests
+import subprocess
+import webbrowser
+import time
+from pathlib import Path
+from datetime import datetime
+from typing import Dict, Any
+from dotenv import load_dotenv
+
+load_dotenv()
+
+# ============================================================================
+# CONFIGURATION
+# ============================================================================
+
+LLM_SERVER = os.getenv("LLM_SERVER_URL", "http://localhost:8000")
+OLLAMA_URL = os.getenv("OLLAMA_URL", "http://localhost:11434")
+LLM_FAST = os.getenv("LLM_FAST_MODEL", "phi")
+LLM_POWER = os.getenv("LLM_POWER_MODEL", "deepseek-coder:33b")
+
+# ============================================================================
+# CODE AGENT CLASS
+# ============================================================================
+
+class CodeAgent:
+    """Agent that modifies files, executes code, and tests in browser"""
+    
+    def __init__(self, work_dir: str = "./workspace"):
+        self.work_dir = Path(work_dir)
+        self.work_dir.mkdir(exist_ok=True)
+        print(f"📁 Workspace: {self.work_dir.absolute()}")
+    
+    # ========================================================================
+    # FILE OPERATIONS
+    # ========================================================================
+    
+    def read_file(self, file_path: str) -> str:
+        """Read file from workspace"""
+        file = self.work_dir / file_path
+        
+        if not file.exists():
+            return f"❌ File not found: {file_path}"
+        
+        try:
+            with open(file, 'r', encoding='utf-8') as f:
+                content = f.read()
+            print(f"✅ Read: {file_path} ({len(content)} bytes)")
+            return content
+        except Exception as e:
+            return f"❌ Error reading {file_path}: {e}"
+    
+    def write_file(self, file_path: str, content: str) -> str:
+        """Write/modify file in workspace"""
+        file = self.work_dir / file_path
+        file.parent.mkdir(parents=True, exist_ok=True)
+        
+        try:
+            with open(file, 'w', encoding='utf-8') as f:
+                f.write(content)
+            print(f"✅ Wrote: {file_path}")
+            return f"File saved: {file_path}"
+        except Exception as e:
+            return f"❌ Error writing {file_path}: {e}"
+    
+    def delete_file(self, file_path: str) -> str:
+        """Delete file from workspace"""
+        file = self.work_dir / file_path
+        
+        if not file.exists():
+            return f"❌ File not found: {file_path}"
+        
+        try:
+            file.unlink()
+            return f"✅ Deleted: {file_path}"
+        except Exception as e:
+            return f"❌ Error deleting {file_path}: {e}"
+    
+    def list_files(self) -> str:
+        """List all files in workspace"""
+        files = list(self.work_dir.rglob("*"))
+        file_list = "\n".join([
+            f"  {f.relative_to(self.work_dir)}" 
+            for f in files if f.is_file()
+        ])
+        return f"📁 Files in workspace:\n{file_list}" if file_list else "📁 No files"
+    
+    # ========================================================================
+    # CODE EXECUTION
+    # ========================================================================
+    
+    def execute_python(self, code: str, file_name: str = "exec.py") -> str:
+        """Execute Python code"""
+        script_path = self.work_dir / file_name
+        
+        try:
+            # Save script
+            with open(script_path, 'w', encoding='utf-8') as f:
+                f.write(code)
+            
+            # Execute
+            result = subprocess.run(
+                ["python3", str(script_path)],
+                capture_output=True,
+                text=True,
+                timeout=30,
+                cwd=self.work_dir
+            )
+            
+            output = result.stdout + result.stderr
+            print(f"🐍 Python executed: {file_name}")
+            return output[:1000]  # Limit output
+        
+        except subprocess.TimeoutExpired:
+            return "❌ Timeout: Script took more than 30 seconds"
+        except Exception as e:
+            return f"❌ Error: {e}"
+    
+    def execute_node(self, code: str, file_name: str = "exec.js") -> str:
+        """Execute Node.js code"""
+        script_path = self.work_dir / file_name
+        
+        try:
+            with open(script_path, 'w', encoding='utf-8') as f:
+                f.write(code)
+            
+            result = subprocess.run(
+                ["node", str(script_path)],
+                capture_output=True,
+                text=True,
+                timeout=30,
+                cwd=self.work_dir
+            )
+            
+            output = result.stdout + result.stderr
+            print(f"📟 Node.js executed: {file_name}")
+            return output[:1000]
+        
+        except FileNotFoundError:
+            return "❌ Node.js not installed"
+        except Exception as e:
+            return f"❌ Error: {e}"
+    
+    def execute_shell(self, command: str) -> str:
+        """Execute shell command (whitelist safe commands only)"""
+        allowed_commands = [
+            'ls', 'pwd', 'mkdir', 'rm', 'cp', 'mv',
+            'cat', 'grep', 'find', 'du', 'df',
+            'git', 'npm', 'pip', 'python'
+        ]
+        
+        cmd_name = command.split()[0] if command else ""
+        if cmd_name not in allowed_commands:
+            return f"❌ Command not allowed: {cmd_name}"
+        
+        try:
+            result = subprocess.run(
+                command.split(),
+                capture_output=True,
+                text=True,
+                timeout=30,
+                cwd=self.work_dir
+            )
+            output = result.stdout + result.stderr
+            return output[:500]
+        except Exception as e:
+            return f"❌ Error: {e}"
+    
+    # ========================================================================
+    # TESTING & BROWSER
+    # ========================================================================
+    
+    def test_html(self, html_file: str) -> str:
+        """Open HTML file in browser for testing"""
+        file_path = self.work_dir / html_file
+        
+        if not file_path.exists():
+            return f"❌ File not found: {html_file}"
+        
+        try:
+            url = f"file:///{file_path.absolute()}".replace("\\", "/")
+            webbrowser.open(url)
+            print(f"🌐 Opened in browser: {html_file}")
+            return f"✅ HTML opened: {html_file}"
+        except Exception as e:
+            return f"❌ Error: {e}"
+    
+    def open_browser(self, url: str) -> str:
+        """Open URL in browser"""
+        try:
+            webbrowser.open(url)
+            return f"✅ Browser opened: {url}"
+        except Exception as e:
+            return f"❌ Error: {e}"
+    
+    # ========================================================================
+    # GIT OPERATIONS
+    # ========================================================================
+    
+    def git_commit(self, message: str = None) -> str:
+        """Auto-commit changes"""
+        try:
+            if not message:
+                message = f"Auto-commit from agent - {datetime.now().isoformat()}"
+            
+            # Git add
+            result = subprocess.run(
+                ['git', 'add', '.'],
+                cwd=self.work_dir,
+                capture_output=True,
+                text=True
+            )
+            
+            if result.returncode != 0:
+                return f"❌ Git add failed: {result.stderr}"
+            
+            # Git commit
+            result = subprocess.run(
+                ['git', 'commit', '-m', message],
+                cwd=self.work_dir,
+                capture_output=True,
+                text=True
+            )
+            
+            if result.returncode != 0:
+                return f"ℹ️ Nothing to commit"
+            
+            return f"✅ Committed: {message}"
+        except Exception as e:
+            return f"❌ Error: {e}"
+    
+    def git_push(self) -> str:
+        """Push to remote repository"""
+        try:
+            result = subprocess.run(
+                ['git', 'push'],
+                cwd=self.work_dir,
+                capture_output=True,
+                text=True,
+                timeout=30
+            )
+            
+            if result.returncode == 0:
+                return "✅ Pushed to remote"
+            else:
+                return f"❌ Push failed: {result.stderr}"
+        except Exception as e:
+            return f"❌ Error: {e}"
+    
+    # ========================================================================
+    # LLM INTERACTION
+    # ========================================================================
+    
+    def call_llm(self, prompt: str, model: str = "phi") -> str:
+        """Call LLM server"""
+        try:
+            response = requests.post(
+                f"{OLLAMA_URL}/api/generate",
+                json={
+                    "model": model,
+                    "prompt": prompt,
+                    "stream": False
+                },
+                timeout=300
+            )
+            
+            if response.status_code == 200:
+                return response.json()["response"]
+            else:
+                return f"❌ LLM Error: {response.status_code}"
+        
+        except requests.exceptions.ConnectionError:
+            return f"❌ Cannot connect to {OLLAMA_URL}"
+        except Exception as e:
+            return f"❌ Error: {e}"
+
+# ============================================================================
+# GET AGENT TOOLS
+# ============================================================================
+
+def get_agent_tools(agent: CodeAgent) -> Dict[str, Any]:
+    """Return dictionary of available tools"""
+    return {
+        "read_file": agent.read_file,
+        "write_file": agent.write_file,
+        "delete_file": agent.delete_file,
+        "execute_python": agent.execute_python,
+        "execute_node": agent.execute_node,
+        "execute_shell": agent.execute_shell,
+        "test_html": agent.test_html,
+        "open_browser": agent.open_browser,
+        "git_commit": agent.git_commit,
+        "git_push": agent.git_push,
+        "list_files": agent.list_files,
+        "call_llm": agent.call_llm,
+    }
+
+# ============================================================================
+# MAIN (for testing)
+# ============================================================================
+
+if __name__ == "__main__":
+    agent = CodeAgent()
+    
+    # Test
+    print("\n🧪 Testing Agent...")
+    print(agent.list_files())
+    
+    # Create test file
+    agent.write_file("test.txt", "Hello from agent!")
+    print(agent.read_file("test.txt"))

backend.py

@@ -0,0 +1,247 @@
+from fastapi import FastAPI, File, UploadFile, Form, HTTPException
+from fastapi.responses import StreamingResponse, JSONResponse
+from fastapi.staticfiles import StaticFiles
+from fastapi.cors import CORSMiddleware
+import aiofiles
+import base64
+import requests
+import json
+import os
+from pathlib import Path
+from dotenv import load_dotenv
+import asyncio
+import io
+from PIL import Image
+import uuid
+
+load_dotenv()
+
+app = FastAPI(title="Code Agent Chat")
+
+# CORS para que el frontend acceda
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Config
+LLM_SERVER = os.getenv("LLM_SERVER_URL", "http://localhost:8000")
+UPLOAD_DIR = Path("./uploads")
+WORKSPACE_DIR = Path("./workspace")
+UPLOAD_DIR.mkdir(exist_ok=True)
+WORKSPACE_DIR.mkdir(exist_ok=True)
+
+# Importar agent
+from agent import CodeAgent, get_agent_tools
+
+agent = CodeAgent(work_dir=str(WORKSPACE_DIR))
+
+# ============================================================================
+# ENDPOINTS
+# ============================================================================
+
+@app.get("/api/health")
+async def health():
+    """Verificar que el servidor está vivo"""
+    return {"status": "ok", "workspace": str(WORKSPACE_DIR.absolute())}
+
+@app.get("/api/files")
+async def list_files():
+    """Listar archivos en workspace"""
+    files = []
+    for f in WORKSPACE_DIR.rglob("*"):
+        if f.is_file():
+            files.append({
+                "name": str(f.relative_to(WORKSPACE_DIR)),
+                "size": f.stat().st_size,
+                "type": f.suffix
+            })
+    return {"files": files}
+
+@app.get("/api/file/{file_path:path}")
+async def read_file(file_path: str):
+    """Leer contenido de archivo"""
+    file = WORKSPACE_DIR / file_path
+    
+    if not file.exists():
+        raise HTTPException(status_code=404, detail="Archivo no encontrado")
+    
+    try:
+        with open(file, 'r') as f:
+            content = f.read()
+        return {"content": content, "name": file_path}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+@app.post("/api/chat")
+async def chat(
+    message: str = Form(...),
+    image: UploadFile = File(None),
+    model: str = Form("phi")
+):
+    """
+    Endpoint principal de chat
+    - message: tu prompt
+    - image: imagen adjunta (opcional)
+    - model: "phi" (rápido) o "deepseek-coder:33b" (potente)
+    """
+    
+    # Procesar imagen si hay
+    image_data = None
+    image_base64 = None
+    
+    if image:
+        try:
+            contents = await image.read()
+            image_base64 = base64.b64encode(contents).decode()
+            
+            # Guardar imagen
+            image_path = UPLOAD_DIR / f"{uuid.uuid4()}.png"
+            with open(image_path, 'wb') as f:
+                f.write(contents)
+            
+            image_data = f"[Imagen adjunta: {image.filename}]"
+            print(f"📸 Imagen procesada: {image.filename}")
+        
+        except Exception as e:
+            raise HTTPException(status_code=400, detail=f"Error procesando imagen: {e}")
+    
+    # Construir prompt para el LLM
+    full_prompt = f"""
+{message}
+
+{f"Contexto de imagen: {image_data}" if image_data else ""}
+
+Tienes acceso a estas herramientas:
+- read_file(file_path): Leer archivo
+- write_file(file_path, content): Escribir archivo
+- execute_python(code): Ejecutar Python
+- execute_node(code): Ejecutar Node.js
+- test_html(file_name): Servir y testear HTML
+- open_browser(url): Abrir navegador
+- list_files(): Listar archivos
+
+Responde en JSON con tu pensamiento y la acción a ejecutar.
+Ejemplo:
+{{"thought": "Voy a leer el archivo", "action": "read_file", "args": {{"file_path": "app.js"}}}}
+"""
+    
+    # Llamar LLM
+    try:
+        response = requests.post(
+            f"{LLM_SERVER}/code",
+            json={
+                "message": full_prompt,
+                "model": model
+            },
+            timeout=300,
+            stream=False
+        )
+        
+        if response.status_code != 200:
+            raise HTTPException(
+                status_code=response.status_code,
+                detail=response.text
+            )
+        
+        llm_response = response.json()["response"]
+        
+        # Parsear JSON y ejecutar acción
+        result = process_agent_response(llm_response)
+        
+        return {
+            "message": message,
+            "image": image.filename if image else None,
+            "model": model,
+            "llm_response": llm_response,
+            "action_result": result,
+            "image_base64": image_base64
+        }
+    
+    except requests.exceptions.ConnectionError:
+        raise HTTPException(
+            status_code=503,
+            detail=f"No se puede conectar a LLM Server: {LLM_SERVER}"
+        )
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+@app.post("/api/chat-stream")
+async def chat_stream(
+    message: str = Form(...),
+    image: UploadFile = File(None),
+    model: str = Form("phi")
+):
+    """Versión streaming (respuestas en tiempo real)"""
+    
+    image_data = None
+    if image:
+        contents = await image.read()
+        image_path = UPLOAD_DIR / f"{uuid.uuid4()}.png"
+        with open(image_path, 'wb') as f:
+            f.write(contents)
+        image_data = f"[Imagen: {image.filename}]"
+    
+    full_prompt = f"{message}\n{image_data if image_data else ''}"
+    
+    async def generate():
+        try:
+            # Streaming desde LLM
+            response = requests.post(
+                f"{LLM_SERVER}/code",
+                json={
+                    "message": full_prompt,
+                    "model": model
+                },
+                timeout=300,
+                stream=True
+            )
+            
+            for line in response.iter_lines():
+                if line:
+                    yield line + b'\n'
+        
+        except Exception as e:
+            yield json.dumps({"error": str(e)}).encode() + b'\n'
+    
+    return StreamingResponse(generate(), media_type="text/event-stream")
+
+def process_agent_response(response: str) -> dict:
+    """Parsear respuesta del LLM y ejecutar acción"""
+    tools = get_agent_tools(agent)
+    
+    try:
+        # Extraer JSON
+        json_start = response.find('{')
+        json_end = response.rfind('}') + 1
+        json_str = response[json_start:json_end]
+        action = json.loads(json_str)
+    except:
+        return {"error": "No se pudo parsear la respuesta", "raw": response[:200]}
+    
+    action_name = action.get("action") or action.get("tool")
+    args = action.get("args", {})
+    
+    if action_name not in tools:
+        return {"error": f"Acción desconocida: {action_name}"}
+    
+    try:
+        result = tools[action_name](**args)
+        return {"success": True, "action": action_name, "result": result}
+    except Exception as e:
+        return {"error": f"Error ejecutando {action_name}: {str(e)}"}
+
+# ============================================================================
+# SERVIR STATIC FILES (UI Frontend)
+# ============================================================================
+
+app.mount("/", StaticFiles(directory="static", html=True), name="static")
+
+if __name__ == "__main__":
+    import uvicorn
+    print(f"🚀 Backend en http://localhost:5000")
+    print(f"📁 Workspace: {WORKSPACE_DIR.absolute()}")
+    uvicorn.run(app, host="0.0.0.0", port=5000)

config/server.json

@@ -0,0 +1,74 @@
+{
+  "server": {
+    "host": "0.0.0.0",
+    "api_port": 8000,
+    "web_port": 5000,
+    "workers": 4,
+    "debug": false,
+    "environment": "production"
+  },
+  "models": {
+    "fast": {
+      "name": "phi",
+      "description": "Phi 2.7B - Rápido",
+      "vram_required_gb": 2,
+      "tokens_per_second": 50
+    },
+    "power": {
+      "name": "deepseek-coder:33b",
+      "description": "DeepSeek 33B - Potente",
+      "vram_required_gb": 18,
+      "tokens_per_second": 6
+    }
+  },
+  "gpu": {
+    "memory_fraction": 0.9,
+    "max_batch_size": 8,
+    "enable_cuda": true,
+    "device_id": 0
+  },
+  "logging": {
+    "level": "INFO",
+    "format": "%(asctime)s - %(name)s - %(levelname)s - %(message)s",
+    "log_dir": "./logs",
+    "max_file_size_mb": 10,
+    "backup_count": 10
+  },
+  "database": {
+    "type": "sqlite",
+    "path": "./data/llm_history.db",
+    "enable_history": true
+  },
+  "security": {
+    "enable_rate_limit": true,
+    "rate_limit_requests": 100,
+    "rate_limit_period_seconds": 60,
+    "require_api_key": false,
+    "allowed_origins": ["*"]
+  },
+  "workspace": {
+    "path": "./workspace",
+    "max_file_size_mb": 100,
+    "allowed_extensions": [".py", ".js", ".html", ".css", ".json", ".txt", ".md"]
+  },
+  "api_endpoints": {
+    "health": "/api/health",
+    "models": "/api/models",
+    "chat": "/api/chat",
+    "files": "/api/files",
+    "metrics": "/api/metrics"
+  },
+  "ollama": {
+    "host": "localhost",
+    "port": 11434,
+    "timeout_seconds": 300,
+    "auto_download_models": true
+  },
+  "features": {
+    "enable_git_commit": true,
+    "enable_code_execution": true,
+    "enable_file_operations": true,
+    "enable_browser_testing": true,
+    "enable_metrics": true
+  }
+}

install_openclaw.sh

@@ -0,0 +1,74 @@
+#!/usr/bin/env bash
+
+set -Eeuo pipefail
+
+readonly INSTALLER_URL="https://openclaw.ai/install.sh"
+readonly SCRIPT_NAME="$(basename "$0")"
+
+show_help() {
+    cat <<EOF
+Uso: $SCRIPT_NAME [--onboard]
+
+Instala OpenClaw usando el instalador oficial.
+
+Opciones:
+  --onboard  Ejecuta el asistente inicial después de instalar.
+  -h, --help Muestra esta ayuda.
+EOF
+}
+
+onboard=false
+
+while [[ $# -gt 0 ]]; do
+    case "$1" in
+        --onboard)
+            onboard=true
+            shift
+            ;;
+        -h|--help)
+            show_help
+            exit 0
+            ;;
+        *)
+            printf 'Error: opción desconocida: %s\n\n' "$1" >&2
+            show_help >&2
+            exit 2
+            ;;
+    esac
+done
+
+if [[ "$(uname -s)" != "Linux" ]]; then
+    printf 'Error: este script está pensado para Linux/WSL2.\n' >&2
+    exit 1
+fi
+
+if ! command -v curl >/dev/null 2>&1; then
+    printf 'Error: curl es obligatorio. Instálalo con: sudo apt-get update && sudo apt-get install -y curl\n' >&2
+    exit 1
+fi
+
+installer_file="$(mktemp)"
+cleanup() {
+    rm -f "$installer_file"
+}
+trap cleanup EXIT
+
+printf 'Descargando el instalador oficial de OpenClaw...\n'
+curl --fail --silent --show-error --location \
+    --proto '=https' --tlsv1.2 \
+    "$INSTALLER_URL" -o "$installer_file"
+
+if [[ "$onboard" == true ]]; then
+    bash "$installer_file"
+else
+    bash "$installer_file" --no-onboard
+fi
+
+printf '\nOpenClaw se ha instalado. Comprueba la instalación con:\n'
+printf '  openclaw --version\n'
+printf '  openclaw doctor\n'
+
+if [[ "$onboard" == false ]]; then
+    printf '\nPara configurar OpenClaw más adelante, ejecuta:\n'
+    printf '  openclaw onboard --install-daemon\n'
+fi

requirements.txt

@@ -0,0 +1,22 @@
+# API & Web Framework
+fastapi==0.104.1
+uvicorn[standard]==0.24.0
+python-multipart==0.0.6
+aiofiles==23.2.1
+
+# LLM & AI
+langchain==0.1.0
+requests==2.31.0
+
+# Data Processing
+pillow==10.1.0
+pydantic==2.5.0
+
+# Configuration
+python-dotenv==1.0.0
+
+# Development (opcional)
+pytest==7.4.3
+pytest-asyncio==0.21.1
+black==23.12.0
+flake8==6.1.0

scripts/backup.sh

@@ -0,0 +1,54 @@
+#!/bin/bash
+
+################################################################################
+# BACKUP SCRIPT
+# Realiza backup de modelos y configuración
+################################################################################
+
+set -e
+
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+BLUE='\033[0;34m'
+NC='\033[0m'
+
+SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
+PROJECT_DIR="$(dirname "$SCRIPT_DIR")"
+BACKUP_DIR="$PROJECT_DIR/backups"
+TIMESTAMP=$(date +%Y%m%d_%H%M%S)
+
+log() { echo -e "${BLUE}[$(date '+%H:%M:%S')]${NC} $@"; }
+log_success() { echo -e "${GREEN}✅ $@${NC}"; }
+
+log "Iniciando backup..."
+
+mkdir -p "$BACKUP_DIR"
+
+# Backup configuración
+log "Haciendo backup de configuración..."
+tar -czf "$BACKUP_DIR/config_$TIMESTAMP.tar.gz" \
+    -C "$PROJECT_DIR" \
+    config/ \
+    .env \
+    requirements.txt \
+    2>/dev/null || true
+
+# Backup modelos (OPCIONAL - muy grandes)
+read -p "¿Hacer backup de modelos Ollama? (y/n - muy grande, ~20GB): " -n 1 -r
+echo
+if [[ $REPLY =~ ^[Yy]$ ]]; then
+    log "Haciendo backup de modelos (esto tardará)..."
+    tar -czf "$BACKUP_DIR/models_$TIMESTAMP.tar.gz" \
+        /home/charle/.ollama/models \
+        2>/dev/null || true
+fi
+
+# Backup workspace
+log "Haciendo backup de workspace..."
+tar -czf "$BACKUP_DIR/workspace_$TIMESTAMP.tar.gz" \
+    -C "$PROJECT_DIR" \
+    workspace/ \
+    2>/dev/null || true
+
+log_success "Backups completados en: $BACKUP_DIR"
+ls -lh "$BACKUP_DIR"

scripts/health-check.sh

@@ -0,0 +1,88 @@
+#!/bin/bash
+
+################################################################################
+# HEALTH CHECK SCRIPT
+# Verifica que todo funcione correctamente
+################################################################################
+
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m'
+
+FAILURES=0
+
+check() {
+    local name=$1
+    local cmd=$2
+    local expected=$3
+
+    echo -n "Verificando $name... "
+
+    if eval "$cmd" &>/dev/null; then
+        echo -e "${GREEN}✅${NC}"
+        return 0
+    else
+        echo -e "${RED}❌${NC}"
+        ((FAILURES++))
+        return 1
+    fi
+}
+
+echo -e "${BLUE}========================================${NC}"
+echo -e "${BLUE}  LLM SERVER HEALTH CHECK${NC}"
+echo -e "${BLUE}========================================${NC}\n"
+
+# Sistema
+echo -e "${YELLOW}Sistema:${NC}"
+check "Ubuntu 22.04" "grep -q '22.04' /etc/os-release"
+check "Internet" "ping -c 1 8.8.8.8"
+
+# Drivers & GPU
+echo -e "\n${YELLOW}GPU & Drivers:${NC}"
+check "NVIDIA Driver" "command -v nvidia-smi"
+check "CUDA 12.3" "command -v nvcc"
+check "RTX 3090" "nvidia-smi | grep -q 'RTX 3090'"
+check "cuDNN" "ldconfig -p | grep -q cudnn"
+
+# Software
+echo -e "\n${YELLOW}Software:${NC}"
+check "Python 3.11" "python3.11 --version"
+check "Docker" "command -v docker"
+check "Git" "command -v git"
+
+# Services
+echo -e "\n${YELLOW}Servicios:${NC}"
+check "Ollama service" "sudo systemctl is-active ollama"
+check "API service" "sudo systemctl is-active llm-api"
+
+# Conectividad
+echo -e "\n${YELLOW}API Connectivity:${NC}"
+check "Ollama API" "curl -s http://localhost:11434/api/tags"
+check "FastAPI" "curl -s http://localhost:8000/health"
+
+# Modelos
+echo -e "\n${YELLOW}Modelos Ollama:${NC}"
+check "Phi disponible" "curl -s http://localhost:11434/api/tags | grep -q 'phi'"
+check "DeepSeek disponible" "curl -s http://localhost:11434/api/tags | grep -q 'deepseek'"
+
+# Disk Space
+echo -e "\n${YELLOW}Espacio en Disco:${NC}"
+root_usage=$(df / | awk 'NR==2 {print int($5)}')
+if [ "$root_usage" -lt 90 ]; then
+    echo -e "Uso de disco (root): ${GREEN}${root_usage}%${NC}"
+else
+    echo -e "Uso de disco (root): ${RED}${root_usage}%${NC}"
+    ((FAILURES++))
+fi
+
+# Summary
+echo -e "\n${BLUE}========================================${NC}"
+if [ "$FAILURES" -eq 0 ]; then
+    echo -e "${GREEN}✅ TODOS LOS CHECKS PASARON${NC}"
+    exit 0
+else
+    echo -e "${RED}❌ $FAILURES CHECKS FALLARON${NC}"
+    exit 1
+fi

scripts/install.sh

@@ -0,0 +1,534 @@
+#!/bin/bash
+
+################################################################################
+# LLM SERVER DEPLOYMENT SCRIPT
+# Ubuntu 22.04 LTS + ASUS H510M + RTX 3090
+# 
+# Uso: bash scripts/install.sh [--quick] [--gpu-only]
+# 
+# Opciones:
+#   --quick        Salta verificaciones lentas
+#   --gpu-only     Solo instala GPU drivers (para re-install)
+################################################################################
+
+set -e  # Exit si hay error
+
+# Colores
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# Variables
+SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
+PROJECT_DIR="$(dirname "$SCRIPT_DIR")"
+WORKSPACE_DIR="$PROJECT_DIR/workspace"
+VENV_DIR="$PROJECT_DIR/venv"
+LOG_FILE="$PROJECT_DIR/logs/install.log"
+
+# Crear directorio de logs
+mkdir -p "$PROJECT_DIR/logs"
+
+# Function: Log con timestamp
+log() {
+    local level=$1
+    shift
+    local message="$@"
+    local timestamp=$(date '+%Y-%m-%d %H:%M:%S')
+    echo -e "${BLUE}[$timestamp]${NC} ${level}: ${message}" | tee -a "$LOG_FILE"
+}
+
+# Function: Log success
+log_success() {
+    echo -e "${GREEN}✅ $@${NC}" | tee -a "$LOG_FILE"
+}
+
+# Function: Log error
+log_error() {
+    echo -e "${RED}❌ $@${NC}" | tee -a "$LOG_FILE"
+}
+
+# Function: Log warning
+log_warning() {
+    echo -e "${YELLOW}⚠️  $@${NC}" | tee -a "$LOG_FILE"
+}
+
+# Function: Check command exists
+command_exists() {
+    command -v "$1" >/dev/null 2>&1
+}
+
+# Function: Check if running as root
+check_root() {
+    if [[ $EUID -ne 0 ]]; then
+        log_error "Este script debe ejecutarse con sudo"
+        exit 1
+    fi
+}
+
+################################################################################
+# MAIN INSTALLATION
+################################################################################
+
+main() {
+    local quick_mode=false
+    local gpu_only=false
+
+    # Parse arguments
+    while [[ $# -gt 0 ]]; do
+        case $1 in
+            --quick)
+                quick_mode=true
+                shift
+                ;;
+            --gpu-only)
+                gpu_only=true
+                shift
+                ;;
+            *)
+                log_error "Opción desconocida: $1"
+                usage
+                exit 1
+                ;;
+        esac
+    done
+
+    log "LOG" "=========================================="
+    log "LOG" "LLM Server Installation"
+    log "LOG" "=========================================="
+    log "LOG" "Project Dir: $PROJECT_DIR"
+    log "LOG" "Workspace: $WORKSPACE_DIR"
+    log "LOG" "Quick Mode: $quick_mode"
+    log "LOG" "GPU Only: $gpu_only"
+    log "LOG" "=========================================="
+
+    # Checks iniciales
+    check_root
+    check_os
+    check_hardware
+
+    if [ "$gpu_only" = false ]; then
+        install_dependencies
+        install_docker
+    fi
+
+    install_nvidia_drivers
+    install_cuda_toolkit
+    install_ollama
+    setup_python_venv
+    create_services
+
+    if [ "$quick_mode" = false ]; then
+        download_models
+    fi
+
+    setup_directories
+    generate_config
+
+    log_success "=========================================="
+    log_success "✨ INSTALACIÓN COMPLETADA"
+    log_success "=========================================="
+    print_next_steps
+}
+
+################################################################################
+# FUNCIONES AUXILIARES
+################################################################################
+
+check_os() {
+    log "LOG" "Verificando Sistema Operativo..."
+    
+    if [ ! -f /etc/os-release ]; then
+        log_error "No se pudo detectar el SO"
+        exit 1
+    fi
+
+    . /etc/os-release
+    if [[ "$ID" != "ubuntu" ]]; then
+        log_error "Este script solo soporta Ubuntu"
+        exit 1
+    fi
+
+    if [[ "$VERSION_ID" != "22.04" ]]; then
+        log_warning "Se recomienda Ubuntu 22.04, detectado: $VERSION_ID"
+    fi
+
+    log_success "Ubuntu $VERSION_ID detectado"
+}
+
+check_hardware() {
+    log "LOG" "Verificando Hardware..."
+
+    # Check GPU
+    if ! command_exists nvidia-smi; then
+        log_warning "nvidia-smi no disponible aún (se instalará)"
+    else
+        gpu_info=$(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader)
+        log_success "GPU detectada: $gpu_info"
+    fi
+
+    # Check CPU
+    cpu_count=$(nproc)
+    log_success "CPU: $cpu_count cores"
+
+    # Check RAM
+    ram_gb=$(free -h | awk '/^Mem:/ {print $2}')
+    log_success "RAM: $ram_gb"
+
+    # Check Disk
+    disk_info=$(df -h / | awk 'NR==2 {print $2}')
+    log_success "Disco: $disk_info disponible"
+}
+
+install_dependencies() {
+    log "LOG" "Instalando dependencias del sistema..."
+
+    apt update
+    apt install -y \
+        build-essential \
+        git \
+        wget \
+        curl \
+        htop \
+        nano \
+        openssh-server \
+        python3.11 \
+        python3.11-venv \
+        python3.11-dev \
+        pkg-config \
+        libssl-dev \
+        libffi-dev
+
+    log_success "Dependencias instaladas"
+}
+
+install_docker() {
+    log "LOG" "Instalando Docker..."
+
+    if command_exists docker; then
+        log_success "Docker ya está instalado"
+        return
+    fi
+
+    curl -fsSL https://get.docker.com -o /tmp/get-docker.sh
+    sh /tmp/get-docker.sh
+
+    # Agregar usuario al grupo docker
+    if id "charle" &>/dev/null; then
+        usermod -aG docker charle
+        log_success "Usuario 'charle' agregado al grupo docker"
+    fi
+
+    systemctl start docker
+    systemctl enable docker
+
+    log_success "Docker instalado"
+}
+
+install_nvidia_drivers() {
+    log "LOG" "Instalando NVIDIA Drivers..."
+
+    if command_exists nvidia-smi; then
+        current_driver=$(nvidia-smi --query-gpu=driver_version --format=csv,noheader | head -1)
+        log_success "Driver NVIDIA $current_driver ya instalado"
+        return
+    fi
+
+    # Agregar repositorio NVIDIA
+    apt-key adv --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC 2>&1 | grep -v "Warning" || true
+
+    add-apt-repository "deb http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/ /" 2>&1 | grep -v "already" || true
+
+    apt update
+    apt install -y cuda-drivers
+
+    log_success "NVIDIA Drivers instalados"
+
+    log_warning "Se recomienda reiniciar: sudo reboot"
+}
+
+install_cuda_toolkit() {
+    log "LOG" "Instalando CUDA Toolkit 12.3..."
+
+    if [ -d "/usr/local/cuda-12.3" ]; then
+        log_success "CUDA 12.3 ya está instalado"
+        return
+    fi
+
+    log "LOG" "Descargando CUDA 12.3.0..."
+    cd /tmp
+    wget -q https://developer.download.nvidia.com/compute/cuda/12.3.0/local_installers/cuda_12.3.0_545.23.06_linux.run \
+        -O cuda_12.3.0_545.23.06_linux.run
+
+    log "LOG" "Instalando CUDA (esto puede tardar 10-15 minutos)..."
+    chmod +x cuda_12.3.0_545.23.06_linux.run
+    ./cuda_12.3.0_545.23.06_linux.run --silent --driver=false --toolkit
+
+    # Configurar PATH
+    if ! grep -q "cuda-12.3" /root/.bashrc; then
+        cat >> /root/.bashrc << 'EOF'
+
+# CUDA 12.3
+export PATH=/usr/local/cuda-12.3/bin:$PATH
+export LD_LIBRARY_PATH=/usr/local/cuda-12.3/lib64:$LD_LIBRARY_PATH
+EOF
+    fi
+
+    source /root/.bashrc
+
+    # Verificar
+    if command_exists nvcc; then
+        cuda_version=$(nvcc --version | grep "release" | awk '{print $5}')
+        log_success "CUDA $cuda_version instalado"
+    fi
+
+    # Install cuDNN
+    log "LOG" "Instalando cuDNN..."
+    apt install -y libcudnn8
+
+    log_success "CUDA Toolkit instalado"
+}
+
+install_ollama() {
+    log "LOG" "Instalando Ollama..."
+
+    if command_exists ollama; then
+        log_success "Ollama ya está instalado"
+    else
+        curl https://ollama.ai/install.sh | sh
+        log_success "Ollama instalado"
+    fi
+
+    # Configurar servicio systemd
+    log "LOG" "Configurando servicio Ollama..."
+
+    mkdir -p /etc/systemd/system
+
+    cat > /etc/systemd/system/ollama.service << 'EOF'
+[Unit]
+Description=Ollama
+After=network-online.target
+
+[Service]
+ExecStart=/usr/local/bin/ollama serve
+User=charle
+Group=charle
+Restart=always
+RestartSec=3
+Environment="OLLAMA_HOST=0.0.0.0:11434"
+Environment="OLLAMA_MODELS=/home/charle/.ollama/models"
+Environment="OLLAMA_NUM_GPU=1"
+
+[Install]
+WantedBy=default.target
+EOF
+
+    systemctl daemon-reload
+    systemctl enable ollama
+    systemctl restart ollama
+
+    sleep 2
+
+    if systemctl is-active --quiet ollama; then
+        log_success "Servicio Ollama activo"
+    else
+        log_error "Error iniciando Ollama"
+    fi
+}
+
+setup_python_venv() {
+    log "LOG" "Creando Python Virtual Environment..."
+
+    if [ -d "$VENV_DIR" ]; then
+        log_success "VEnv ya existe"
+        return
+    fi
+
+    python3.11 -m venv "$VENV_DIR"
+
+    source "$VENV_DIR/bin/activate"
+
+    pip install --upgrade pip setuptools wheel
+
+    pip install \
+        fastapi \
+        uvicorn \
+        python-multipart \
+        aiofiles \
+        pillow \
+        python-dotenv \
+        requests \
+        langchain \
+        pydantic
+
+    log_success "Python VEnv configurado"
+}
+
+create_services() {
+    log "LOG" "Creando systemd services..."
+
+    # API Service
+    cat > /etc/systemd/system/llm-api.service << EOF
+[Unit]
+Description=LLM API Server
+After=network.target ollama.service
+
+[Service]
+Type=simple
+User=charle
+WorkingDirectory=$PROJECT_DIR
+Environment="PATH=$VENV_DIR/bin"
+Environment="OLLAMA_URL=http://localhost:11434"
+ExecStart=$VENV_DIR/bin/python $PROJECT_DIR/backend.py
+Restart=always
+RestartSec=10
+
+[Install]
+WantedBy=multi-user.target
+EOF
+
+    systemctl daemon-reload
+    systemctl enable llm-api
+
+    log_success "Systemd services creados"
+}
+
+download_models() {
+    log "LOG" "Descargando modelos (esto puede tardar 20-30 minutos)..."
+    log_warning "Esto es OPCIONAL. Presiona Ctrl+C para cancelar."
+
+    sleep 5
+
+    source "$VENV_DIR/bin/activate"
+
+    # Esperar a que Ollama esté listo
+    for i in {1..30}; do
+        if curl -s http://localhost:11434/api/tags > /dev/null; then
+            log_success "Ollama está listo"
+            break
+        fi
+        log "LOG" "Esperando Ollama... ($i/30)"
+        sleep 2
+    done
+
+    log "LOG" "Descargando Phi..."
+    ollama pull phi:latest
+
+    log "LOG" "Descargando DeepSeek Coder 33B..."
+    ollama pull deepseek-coder:33b
+
+    log_success "Modelos descargados"
+}
+
+setup_directories() {
+    log "LOG" "Creando directorios..."
+
+    mkdir -p "$WORKSPACE_DIR"
+    mkdir -p "$PROJECT_DIR/uploads"
+    mkdir -p "$PROJECT_DIR/logs"
+
+    # Cambiar permisos
+    chown -R charle:charle "$PROJECT_DIR"
+    chmod -R 755 "$PROJECT_DIR"
+
+    log_success "Directorios creados"
+}
+
+generate_config() {
+    log "LOG" "Generando archivos de configuración..."
+
+    # .env file
+    if [ ! -f "$PROJECT_DIR/.env" ]; then
+        cat > "$PROJECT_DIR/.env" << EOF
+# LLM Server Configuration
+LLM_SERVER_URL=http://localhost:8000
+LLM_FAST_MODEL=phi
+LLM_POWER_MODEL=deepseek-coder:33b
+OLLAMA_URL=http://localhost:11434
+WORKSPACE_DIR=$WORKSPACE_DIR
+
+# Server
+HOST=0.0.0.0
+API_PORT=8000
+WEB_PORT=5000
+
+# Logging
+LOG_LEVEL=INFO
+EOF
+        log_success ".env creado"
+    fi
+
+    # Config JSON
+    if [ ! -f "$PROJECT_DIR/config/server.json" ]; then
+        mkdir -p "$PROJECT_DIR/config"
+        cat > "$PROJECT_DIR/config/server.json" << 'EOF'
+{
+  "server": {
+    "host": "0.0.0.0",
+    "api_port": 8000,
+    "web_port": 5000,
+    "workers": 4
+  },
+  "models": {
+    "fast": "phi",
+    "power": "deepseek-coder:33b"
+  },
+  "gpu": {
+    "memory_fraction": 0.9,
+    "max_batch_size": 8
+  },
+  "logging": {
+    "level": "INFO",
+    "format": "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
+  }
+}
+EOF
+        log_success "config/server.json creado"
+    fi
+}
+
+print_next_steps() {
+    cat << EOF
+
+${BLUE}=========================================${NC}
+${GREEN}✨ PRÓXIMOS PASOS:${NC}
+${BLUE}=========================================${NC}
+
+1. ${YELLOW}Verificar instalación:${NC}
+   sudo systemctl status ollama
+   sudo systemctl status llm-api
+
+2. ${YELLOW}Ver logs:${NC}
+   sudo journalctl -u ollama -f
+   sudo journalctl -u llm-api -f
+
+3. ${YELLOW}Iniciar servicios:${NC}
+   sudo systemctl start ollama
+   sudo systemctl start llm-api
+
+4. ${YELLOW}Acceder a la API:${NC}
+   curl http://localhost:8000/health
+
+5. ${YELLOW}Descargar modelos (si no se descargaron):${NC}
+   source $VENV_DIR/bin/activate
+   ollama pull phi:latest
+   ollama pull deepseek-coder:33b
+
+6. ${YELLOW}Ver estado en tiempo real:${NC}
+   watch -n 1 nvidia-smi
+
+${BLUE}=========================================${NC}
+${GREEN}📁 Archivos importantes:${NC}
+${BLUE}=========================================${NC}
+   Config: $PROJECT_DIR/config/server.json
+   .env: $PROJECT_DIR/.env
+   Logs: $PROJECT_DIR/logs/
+   Workspace: $WORKSPACE_DIR/
+
+${BLUE}=========================================${NC}
+EOF
+}
+
+# Ejecutar main
+main "$@"

scripts/monitor.sh

@@ -0,0 +1,100 @@
+#!/bin/bash
+
+################################################################################
+# MONITOR SCRIPT
+# Monitorea servicios y hardware en tiempo real
+################################################################################
+
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m'
+
+clear_screen() {
+    clear
+}
+
+print_header() {
+    echo -e "${BLUE}=================================================${NC}"
+    echo -e "${BLUE}   LLM SERVER MONITOR - $(date '+%Y-%m-%d %H:%M:%S')${NC}"
+    echo -e "${BLUE}=================================================${NC}"
+}
+
+print_section() {
+    echo -e "\n${YELLOW}>>> $1${NC}"
+}
+
+check_service() {
+    local service=$1
+    if sudo systemctl is-active --quiet "$service"; then
+        echo -e "${GREEN}✅ $service - ACTIVO${NC}"
+        sudo systemctl status "$service" --no-pager | grep -E "(Active|ExecStart)" | sed 's/^/   /'
+    else
+        echo -e "${RED}❌ $service - INACTIVO${NC}"
+    fi
+}
+
+get_ip() {
+    hostname -I | awk '{print $1}'
+}
+
+main() {
+    while true; do
+        clear_screen
+        print_header
+
+        # System Info
+        print_section "SISTEMA"
+        echo "Hostname: $(hostname)"
+        echo "IP: $(get_ip)"
+        echo "Uptime: $(uptime -p)"
+        echo "Usuarios conectados: $(who | wc -l)"
+
+        # CPU & RAM
+        print_section "CPU & MEMORIA"
+        free -h | awk 'NR==1 {print ""; print $0} NR==2 {print $0}'
+        echo ""
+        top -bn1 | head -3 | tail -1
+
+        # Disk
+        print_section "DISCO"
+        df -h / | awk 'NR==2 {printf "Root: %s used / %s total (%.1f%%)\n", $3, $2, ($3/$2)*100}'
+
+        # GPU
+        print_section "GPU - NVIDIA RTX 3090"
+        nvidia-smi --query-gpu=index,name,driver_version,memory.used,memory.total,temperature.gpu,utilization.gpu \
+            --format=csv,noheader | while read line; do
+            echo "  $line"
+        done
+
+        # Services
+        print_section "SERVICIOS"
+        check_service "ollama"
+        echo ""
+        check_service "llm-api"
+
+        # Network
+        print_section "RED"
+        echo "API (port 8000): $(curl -s http://localhost:8000/health | jq '.' 2>/dev/null || echo 'NO RESPONDE')"
+        echo "Ollama (port 11434): $(curl -s http://localhost:11434/api/tags | jq '.models | length' 2>/dev/null || echo '0') modelos"
+
+        # Logs recientes
+        print_section "ÚLTIMOS ERRORES (últimas 5 líneas)"
+        echo "Ollama:"
+        sudo journalctl -u ollama -n 3 --no-pager | sed 's/^/  /'
+        echo ""
+        echo "API:"
+        sudo journalctl -u llm-api -n 3 --no-pager | sed 's/^/  /'
+
+        # Footer
+        echo ""
+        echo -e "${BLUE}=================================================${NC}"
+        echo "Presiona Ctrl+C para salir | Se actualiza cada 10 segundos"
+        echo -e "${BLUE}=================================================${NC}"
+
+        sleep 10
+    done
+}
+
+main

scripts/update.sh

@@ -0,0 +1,53 @@
+#!/bin/bash
+
+################################################################################
+# UPDATE SCRIPT
+# Actualiza código, modelos y dependencias
+################################################################################
+
+set -e
+
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m'
+
+SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
+PROJECT_DIR="$(dirname "$SCRIPT_DIR")"
+VENV_DIR="$PROJECT_DIR/venv"
+
+log() { echo -e "${BLUE}[$(date '+%H:%M:%S')]${NC} $@"; }
+log_success() { echo -e "${GREEN}✅ $@${NC}"; }
+log_error() { echo -e "${RED}❌ $@${NC}"; exit 1; }
+
+log "Actualizando LLM Server..."
+
+# Pull latest from git
+log "Descargando cambios de git..."
+cd "$PROJECT_DIR"
+git pull origin main || log "Git pull completado con warnings"
+
+# Update Python dependencies
+log "Actualizando dependencias Python..."
+source "$VENV_DIR/bin/activate"
+pip install --upgrade pip
+pip install -r requirements.txt --upgrade
+
+# Update Ollama models (opcional)
+read -p "¿Actualizar modelos Ollama? (y/n): " -n 1 -r
+echo
+if [[ $REPLY =~ ^[Yy]$ ]]; then
+    log "Actualizando Phi..."
+    ollama pull phi:latest
+    
+    log "Actualizando DeepSeek..."
+    ollama pull deepseek-coder:33b
+fi
+
+# Restart services
+log "Reiniciando servicios..."
+sudo systemctl restart ollama
+sudo systemctl restart llm-api
+
+log_success "Actualización completada"

static/index.html

@@ -0,0 +1,549 @@
+<!DOCTYPE html>
+<html lang="es">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Code Agent Chat</title>
+    <style>
+        * {
+            margin: 0;
+            padding: 0;
+            box-sizing: border-box;
+        }
+
+        body {
+            font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif;
+            background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+            height: 100vh;
+            display: flex;
+        }
+
+        .container {
+            display: flex;
+            width: 100%;
+            gap: 20px;
+            padding: 20px;
+        }
+
+        /* Chat Panel */
+        .chat-panel {
+            flex: 1;
+            display: flex;
+            flex-direction: column;
+            background: white;
+            border-radius: 12px;
+            box-shadow: 0 8px 32px rgba(0,0,0,0.1);
+            overflow: hidden;
+        }
+
+        .chat-header {
+            background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+            color: white;
+            padding: 20px;
+            font-size: 20px;
+            font-weight: bold;
+        }
+
+        .chat-messages {
+            flex: 1;
+            overflow-y: auto;
+            padding: 20px;
+            display: flex;
+            flex-direction: column;
+            gap: 15px;
+        }
+
+        .message {
+            display: flex;
+            gap: 10px;
+            animation: slideIn 0.3s ease-out;
+        }
+
+        @keyframes slideIn {
+            from { opacity: 0; transform: translateY(10px); }
+            to { opacity: 1; transform: translateY(0); }
+        }
+
+        .message.user {
+            justify-content: flex-end;
+        }
+
+        .message-content {
+            max-width: 70%;
+            padding: 12px 16px;
+            border-radius: 12px;
+            word-wrap: break-word;
+        }
+
+        .message.user .message-content {
+            background: #667eea;
+            color: white;
+            border-bottom-right-radius: 4px;
+        }
+
+        .message.assistant .message-content {
+            background: #f0f0f0;
+            color: #333;
+            border-bottom-left-radius: 4px;
+        }
+
+        .message-image {
+            max-width: 200px;
+            border-radius: 8px;
+            margin-top: 8px;
+        }
+
+        .message-meta {
+            font-size: 12px;
+            color: #999;
+            margin-top: 4px;
+        }
+
+        /* Input Area */
+        .input-area {
+            border-top: 1px solid #e0e0e0;
+            padding: 20px;
+            background: #fafafa;
+        }
+
+        .input-container {
+            display: flex;
+            gap: 10px;
+            margin-bottom: 10px;
+        }
+
+        textarea {
+            flex: 1;
+            padding: 12px;
+            border: 1px solid #ddd;
+            border-radius: 8px;
+            resize: vertical;
+            min-height: 60px;
+            max-height: 120px;
+            font-family: inherit;
+            font-size: 14px;
+        }
+
+        textarea:focus {
+            outline: none;
+            border-color: #667eea;
+            box-shadow: 0 0 0 3px rgba(102, 126, 234, 0.1);
+        }
+
+        .button-group {
+            display: flex;
+            gap: 10px;
+        }
+
+        button {
+            padding: 10px 20px;
+            border: none;
+            border-radius: 8px;
+            cursor: pointer;
+            font-weight: 600;
+            transition: all 0.2s;
+        }
+
+        .send-btn {
+            background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+            color: white;
+            flex: 1;
+        }
+
+        .send-btn:hover {
+            transform: translateY(-2px);
+            box-shadow: 0 4px 12px rgba(102, 126, 234, 0.4);
+        }
+
+        .send-btn:disabled {
+            opacity: 0.5;
+            cursor: not-allowed;
+            transform: none;
+        }
+
+        .file-input-wrapper {
+            position: relative;
+            overflow: hidden;
+        }
+
+        .file-input-wrapper input[type="file"] {
+            display: none;
+        }
+
+        .upload-btn {
+            background: #4CAF50;
+            color: white;
+            padding: 10px 20px;
+        }
+
+        .upload-btn:hover {
+            background: #45a049;
+        }
+
+        .selected-image {
+            max-width: 150px;
+            max-height: 150px;
+            border-radius: 8px;
+            margin-top: 10px;
+        }
+
+        .image-indicator {
+            font-size: 12px;
+            color: #666;
+            margin-top: 5px;
+        }
+
+        /* Sidebar */
+        .sidebar {
+            width: 300px;
+            background: white;
+            border-radius: 12px;
+            box-shadow: 0 8px 32px rgba(0,0,0,0.1);
+            display: flex;
+            flex-direction: column;
+            overflow: hidden;
+        }
+
+        .sidebar-header {
+            background: #f5f5f5;
+            padding: 15px;
+            font-weight: bold;
+            border-bottom: 1px solid #e0e0e0;
+        }
+
+        .sidebar-content {
+            flex: 1;
+            overflow-y: auto;
+            padding: 15px;
+        }
+
+        .file-item {
+            padding: 8px 12px;
+            background: #f9f9f9;
+            border-radius: 6px;
+            margin-bottom: 8px;
+            font-size: 13px;
+            cursor: pointer;
+            transition: all 0.2s;
+            border-left: 3px solid transparent;
+        }
+
+        .file-item:hover {
+            background: #f0f0f0;
+            border-left-color: #667eea;
+        }
+
+        .model-selector {
+            padding: 15px;
+            border-top: 1px solid #e0e0e0;
+        }
+
+        select {
+            width: 100%;
+            padding: 8px;
+            border: 1px solid #ddd;
+            border-radius: 6px;
+        }
+
+        .loading {
+            display: inline-block;
+            width: 8px;
+            height: 8px;
+            background: #667eea;
+            border-radius: 50%;
+            animation: pulse 1.5s infinite;
+            margin-right: 5px;
+        }
+
+        @keyframes pulse {
+            0%, 100% { opacity: 1; }
+            50% { opacity: 0.5; }
+        }
+
+        .error {
+            background: #ffebee;
+            color: #c62828;
+            padding: 12px;
+            border-radius: 8px;
+            margin-top: 10px;
+        }
+
+        .success {
+            background: #e8f5e9;
+            color: #2e7d32;
+            padding: 12px;
+            border-radius: 8px;
+            margin-top: 10px;
+        }
+
+        @media (max-width: 768px) {
+            .container {
+                flex-direction: column;
+            }
+
+            .sidebar {
+                width: 100%;
+            }
+
+            .message-content {
+                max-width: 90%;
+            }
+        }
+    </style>
+</head>
+<body>
+    <div class="container">
+        <!-- Chat Panel -->
+        <div class="chat-panel">
+            <div class="chat-header">🤖 Code Agent Chat</div>
+            
+            <div class="chat-messages" id="chatMessages"></div>
+            
+            <div class="input-area">
+                <div class="input-container">
+                    <textarea 
+                        id="messageInput" 
+                        placeholder="Escribe tu directive aquí... (ej: Crea un contador HTML)"
+                        onkeydown="if(event.key==='Enter' && event.ctrlKey) sendMessage();"
+                    ></textarea>
+                </div>
+                
+                <div id="selectedImageContainer" style="display:none;">
+                    <img id="selectedImagePreview" class="selected-image" alt="Imagen seleccionada">
+                    <div class="image-indicator" id="imageIndicator"></div>
+                </div>
+                
+                <div class="button-group">
+                    <div class="file-input-wrapper">
+                        <input type="file" id="imageInput" accept="image/*">
+                        <button class="upload-btn" onclick="document.getElementById('imageInput').click();">
+                            📸 Adjuntar Imagen
+                        </button>
+                    </div>
+                    <button class="send-btn" id="sendBtn" onclick="sendMessage();">
+                        Enviar (Ctrl+Enter)
+                    </button>
+                </div>
+            </div>
+        </div>
+
+        <!-- Sidebar -->
+        <div class="sidebar">
+            <div class="sidebar-header">📁 Workspace</div>
+            
+            <div class="sidebar-content" id="fileList">
+                <p style="color: #999; font-size: 13px;">Cargando archivos...</p>
+            </div>
+            
+            <div class="model-selector">
+                <label style="font-size: 12px; color: #666;">Modelo LLM:</label>
+                <select id="modelSelect" onchange="localStorage.setItem('selectedModel', this.value)">
+                    <option value="phi">⚡ Phi (Rápido)</option>
+                    <option value="deepseek-coder:33b">🧠 DeepSeek (Potente)</option>
+                </select>
+            </div>
+        </div>
+    </div>
+
+    <script>
+        const API_URL = "http://localhost:5000/api";
+        const chatMessagesDiv = document.getElementById("chatMessages");
+        const messageInput = document.getElementById("messageInput");
+        const sendBtn = document.getElementById("sendBtn");
+        const imageInput = document.getElementById("imageInput");
+        const fileList = document.getElementById("fileList");
+        const modelSelect = document.getElementById("modelSelect");
+        const selectedImageContainer = document.getElementById("selectedImageContainer");
+        const selectedImagePreview = document.getElementById("selectedImagePreview");
+        const imageIndicator = document.getElementById("imageIndicator");
+
+        let selectedImage = null;
+
+        // Load saved model
+        const savedModel = localStorage.getItem("selectedModel") || "phi";
+        modelSelect.value = savedModel;
+
+        // Manejo de imagen
+        imageInput.addEventListener("change", (e) => {
+            const file = e.target.files[0];
+            if (file) {
+                selectedImage = file;
+                const reader = new FileReader();
+                reader.onload = (event) => {
+                    selectedImagePreview.src = event.target.result;
+                    imageIndicator.textContent = `✅ ${file.name} seleccionada`;
+                    selectedImageContainer.style.display = "block";
+                };
+                reader.readAsDataURL(file);
+            }
+        });
+
+        // Limpiar imagen
+        function clearImage() {
+            selectedImage = null;
+            imageInput.value = "";
+            selectedImageContainer.style.display = "none";
+        }
+
+        // Enviar mensaje
+        async function sendMessage() {
+            const message = messageInput.value.trim();
+            if (!message) return;
+
+            const model = modelSelect.value;
+
+            // Mostrar mensaje del usuario
+            addMessage(message, "user", selectedImage);
+            messageInput.value = "";
+            clearImage();
+            sendBtn.disabled = true;
+
+            try {
+                // Preparar form data
+                const formData = new FormData();
+                formData.append("message", message);
+                formData.append("model", model);
+                if (selectedImage) {
+                    formData.append("image", selectedImage);
+                }
+
+                // Enviar
+                const response = await fetch(`${API_URL}/chat`, {
+                    method: "POST",
+                    body: formData
+                });
+
+                if (!response.ok) {
+                    const error = await response.json();
+                    addMessage(`❌ Error: ${error.detail}`, "assistant");
+                    return;
+                }
+
+                const data = await response.json();
+
+                // Mostrar respuesta del LLM
+                addMessage(data.llm_response, "assistant", null, model);
+
+                // Mostrar resultado de la acción
+                if (data.action_result) {
+                    const resultText = JSON.stringify(data.action_result, null, 2);
+                    if (data.action_result.success) {
+                        addMessage(`✅ ${data.action_result.result}`, "assistant");
+                    } else {
+                        addMessage(`⚠️ ${data.action_result.error || data.action_result.result}`, "assistant");
+                    }
+                }
+
+                // Recargar archivos
+                loadFiles();
+            } catch (error) {
+                addMessage(`❌ Error: ${error.message}`, "assistant");
+            } finally {
+                sendBtn.disabled = false;
+                messageInput.focus();
+            }
+        }
+
+        // Agregar mensaje al chat
+        function addMessage(text, sender, image = null, model = null) {
+            const messageDiv = document.createElement("div");
+            messageDiv.className = `message ${sender}`;
+
+            const contentDiv = document.createElement("div");
+            contentDiv.className = "message-content";
+
+            // Formatear texto (markdown básico)
+            let formattedText = text
+                .replace(/```([^`]*?)```/g, "<pre><code>$1</code></pre>")
+                .replace(/\*\*(.*?)\*\*/g, "<strong>$1</strong>")
+                .replace(/_(.*?)_/g, "<em>$1</em>");
+
+            contentDiv.innerHTML = formattedText;
+
+            if (image instanceof File) {
+                const reader = new FileReader();
+                reader.onload = (e) => {
+                    const img = document.createElement("img");
+                    img.src = e.target.result;
+                    img.className = "message-image";
+                    contentDiv.appendChild(img);
+                };
+                reader.readAsDataURL(image);
+            }
+
+            messageDiv.appendChild(contentDiv);
+
+            if (model) {
+                const meta = document.createElement("div");
+                meta.className = "message-meta";
+                meta.textContent = `Modelo: ${model}`;
+                messageDiv.appendChild(meta);
+            }
+
+            chatMessagesDiv.appendChild(messageDiv);
+            chatMessagesDiv.scrollTop = chatMessagesDiv.scrollHeight;
+        }
+
+        // Cargar archivos del workspace
+        async function loadFiles() {
+            try {
+                const response = await fetch(`${API_URL}/files`);
+                const data = await response.json();
+
+                fileList.innerHTML = "";
+
+                if (data.files.length === 0) {
+                    fileList.innerHTML = '<p style="color: #999; font-size: 13px;">Sin archivos aún</p>';
+                    return;
+                }
+
+                data.files.forEach(file => {
+                    const item = document.createElement("div");
+                    item.className = "file-item";
+                    
+                    let icon = "📄";
+                    if (file.type === ".html") icon = "🌐";
+                    if (file.type === ".js") icon = "📜";
+                    if (file.type === ".py") icon = "🐍";
+                    if (file.type === ".css") icon = "🎨";
+                    
+                    item.innerHTML = `${icon} <strong>${file.name}</strong><br>
+                                     <span style="font-size: 11px; color: #999;">${(file.size / 1024).toFixed(1)} KB</span>`;
+                    
+                    item.onclick = async () => {
+                        try {
+                            const fileResponse = await fetch(`${API_URL}/file/${file.name}`);
+                            const fileData = await fileResponse.json();
+                            addMessage(`\`\`\`\n${fileData.content}\n\`\`\``, "assistant");
+                        } catch (e) {
+                            addMessage(`Error leyendo ${file.name}`, "assistant");
+                        }
+                    };
+                    
+                    fileList.appendChild(item);
+                });
+            } catch (error) {
+                console.error("Error cargando archivos:", error);
+            }
+        }
+
+        // Cargar archivos al iniciar
+        loadFiles();
+
+        // Auto-refresh cada 10 segundos
+        setInterval(loadFiles, 10000);
+
+        // Enter para enviar
+        messageInput.addEventListener("keydown", (e) => {
+            if (e.key === "Enter" && e.ctrlKey) {
+                sendMessage();
+            }
+        });
+
+        // Mensaje inicial
+        setTimeout(() => {
+            addMessage("👋 Hola! Soy tu agent agentico. Puedes:\n\n- **Pedir que cree código** (HTML, Python, Node.js)\n- **Adjuntar imágenes** (capturas, diseños)\n- **Dar directivas** para modificar tus archivos\n- **Ejecutar y testear** automáticamente en el navegador\n\n¿Qué querés crear?", "assistant");
+        }, 500);
+    </script>
+</body>
+</html>