Tag: agentic-ai

Chat with your Data: Building a File-Aware AI Agent with AWS Bedrock and Chainlit

~ Gonzalo Ayuso ~ Leave a comment

We all know LLMs are powerful, but their true potential is unlocked when they can see your data. While RAG (Retrieval-Augmented Generation) is great for massive knowledge bases, sometimes you just want to drag and drop a file and ask questions about it.

Today we’ll build a “File-Aware” AI agent that can natively understand a wide range of document formats—from PDFs and Excel sheets to Word docs and Markdown files. We’ll use AWS Bedrock with Claude 4.5 Sonnet for the reasoning engine and Chainlit for the conversational UI.

The idea is straightforward: Upload a file, inject it into the model’s context, and let the LLM do the rest. No vector databases, no complex indexing pipelines—just direct context injection for immediate analysis.

The architecture is simple yet effective. We intercept file uploads in the UI, process them into a format the LLM understands, and pass them along with the user’s query.

┌──────────────┐      ┌──────────────┐      ┌────────────────────┐
│   Chainlit   │      │  Orchestrator│      │   AWS Bedrock      │
│      UI      │─────►│    Agent     │─────►│(Claude 4.5 Sonnet) │
└──────┬───────┘      └──────────────┘      └────────────────────┘
       │                      ▲
       │    ┌────────────┐    │
       └───►│ File Proc. │────┘
            │   Logic    │
            └────────────┘

The tech stack includes:

  • AWS Bedrock with Claude 4.5 Sonnet for high-quality reasoning and large context windows.
  • Chainlit for a chat-like interface with native file upload support.
  • Python for the backend logic.

The core challenge is handling different file types and presenting them to the LLM. We support a variety of formats by mapping them to Bedrock’s expected input types.

To enable file uploads in Chainlit, you need to configure the [features.spontaneous_file_upload] section in your .chainlit/config.toml. This is where you define which MIME types are accepted.

[features.spontaneous_file_upload]
    enabled = true
    accept = [
        "application/pdf",
        "text/csv",
        "application/msword",
        "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
        "application/vnd.ms-excel",
        "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
        "text/html",
        "text/plain",
        "text/markdown",
        "text/x-markdown"
    ]
    max_files = 20
    max_size_mb = 500

The main agent loop handles the conversation. It checks for uploaded files, processes them, and constructs the message payload for the LLM. We also include robust error handling to manage context window limits gracefully.
def get_question_from_message(message: cl.Message):
    content_blocks = None
    if message.elements:
        content_blocks = get_content_blocks_from_message(message)

    if content_blocks:
        content_blocks.append({"text": message.content or "Write a summary of the document"})
        question = content_blocks
    else:
        question = message.content

    return question


def get_content_blocks_from_message(message: cl.Message):
    docs = [f for f in message.elements if f.type == "file" and f.mime in MIME_MAP]
    content_blocks = []

    for doc in docs:
        file = Path(doc.path)
        file_bytes = file.read_bytes()
        shutil.rmtree(file.parent)

        content_blocks.append({
            "document": {
                "name": sanitize_filename(doc.name),
                "format": MIME_MAP[doc.mime],
                "source": {"bytes": file_bytes}
            }
        })

    return content_blocks

@cl.on_message
async def handle_message(message: cl.Message):
    task = asyncio.create_task(process_user_task(
        question=get_question_from_message(message),
        debug=DEBUG))
    cl.user_session.set("task", task)
    try:
        await task
    except asyncio.CancelledError:
        logger.info("User task was cancelled.")

This pattern allows for ad-hoc analysis. You don’t need to pre-ingest data. You can:

  1. Analyze Financials: Upload an Excel sheet and ask for trends.
  2. Review Contracts: Upload a PDF and ask for clause summaries.
  3. Debug Code: Upload a source file and ask for a bug fix.
By leveraging the large context window of modern models like Claude 4.5 Sonnet, we can feed entire documents directly into the prompt, providing the model with full visibility without the information loss often associated with RAG chunking.

And that's all. With tools like Chainlit and powerful APIs like AWS Bedrock, we can create robust, multi-modal assistants that integrate seamlessly into our daily workflows.

Full code in my github account.

Building scalable multi-purpose AI agents: Orchestrating Multi-Agent Systems with Strands Agents and Chainlit

~ Gonzalo Ayuso ~ Leave a comment

We can build simple AI agents that handle specific tasks quite easily today. But what about building AI systems that can handle multiple domains effectively? One approach is to create a single monolithic agent that tries to do everything, but this quickly runs into problems of context pollution, maintenance complexity, and scaling limitations. In this article, we’ll show a production-ready pattern for building multi-purpose AI systems using an orchestrator architecture that coordinates domain-specific agents.

The idea is simple: Don’t build one agent to rule them all instead, create specialized agents that excel in their domains and coordinate them through an intelligent orchestrator. The solution is an orchestrator agent that routes requests to specialized sub-agents, each with focused expertise and dedicated tools. Think of it as a smart router that understands intent and delegates accordingly.

That’s the core of the Orchestrator Pattern for multi-agent systems:

User Query → Orchestrator Agent → Specialized Agent(s) → Orchestrator → Response

For our example we have three specialized agents:

  1. Weather Agent: Expert in meteorological data and weather patterns. It uses external weather APIs to fetch historical and current weather data.
  2. Logistics Agent: Specialist in supply chain and shipping operations. Fake logistics data is generated to simulate shipment tracking, route optimization, and delivery performance analysis.
  3. Production Agent: Focused on manufacturing operations and production metrics. Also, fake production data is generated to analyze production KPIs.

That’s the architecture in a nutshell:

┌─────────────────────────────────────────────┐
│          Orchestrator Agent                 │
│  (Routes & Synthesizes)                 │
└────────┬─────────┬─────────┬────────────────┘
         │         │         │
    ┌────▼────┐ ┌──▼─────┐ ┌─▼─────────┐
    │ Weather │ │Logistic│ │Production │
    │  Agent  │ │ Agent  │ │  Agent    │
    └────┬────┘ └──┬─────┘ └┬──────────┘
         │         │        │
    ┌────▼────┐ ┌──▼─────┐ ┌▼──────────┐
    │External │ │Database│ │ Database  │
    │   API   │ │ Tools  │ │  Tools    │
    └─────────┘ └────────┘ └───────────┘

The tech stack includes:

  • AWS Bedrock with Claude 4.5 Sonnet for agent reasoning
  • Strands Agents framework for agent orchestration
  • Chainlit for the conversational UI
  • FastAPI for the async backend
  • PostgreSQL for storing conversation history and domain data

The orchestrator’s job is simple but critical: understand the user’s intent and route to the right specialist(s).

MAIN_SYSTEM_PROMPT = """You are an intelligent orchestrator agent 
responsible for routing user requests to specialized sub-agents 
based on their domain expertise.

## Available Specialized Agents

### 1. Production Agent
**Domain**: Manufacturing operations, production metrics, quality control
**Handles**: Production KPIs, machine performance, downtime analysis

### 2. Logistics Agent
**Domain**: Supply chain, shipping, transportation operations
**Handles**: Shipment tracking, route optimization, delivery performance

### 3. Weather Agent
**Domain**: Meteorological data and weather patterns
**Handles**: Historical weather, atmospheric conditions, climate trends

## Your Decision Process
1. Analyze the request for key terms and domains
2. Determine scope (single vs multi-domain)
3. Route to appropriate agent(s)
4. Synthesize results when multiple agents are involved
"""

The orchestrator receives specialized agents as tools:

def get_orchestrator_tools() -> List[Any]:
    from tools.logistics.agent import logistics_assistant
    from tools.production.agent import production_assistant
    from tools.weather.agent import weather_assistant

    tools = [
        calculator,
        think,
        current_time,
        AgentCoreCodeInterpreter(region=AWS_REGION).code_interpreter,
        logistics_assistant,  # Specialized agent as tool
        production_assistant,  # Specialized agent as tool
        weather_assistant     # Specialized agent as tool
    ]
    return tools

Each specialized agent follows a consistent pattern. Here’s the weather agent:

@tool
@stream_to_step("weather_assistant")
async def weather_assistant(query: str):
    """
    A research assistant specialized in weather topics with streaming support.
    """
    try:
        tools = [
            calculator,
            think,
            current_time,
            AgentCoreCodeInterpreter(region=AWS_REGION).code_interpreter
        ]
        # Domain-specific tools
        tools += WeatherTools(latitude=MY_LATITUDE, longitude=MY_LONGITUDE).get_tools()

        research_agent = get_agent(
            system_prompt=WEATHER_ASSISTANT_PROMPT,
            tools=tools
        )

        async for token in research_agent.stream_async(query):
            yield token

    except Exception as e:
        yield f"Error in research assistant: {str(e)}"

Each agent has access to domain-specific tools. For example, the weather agent uses external APIs:

class WeatherTools:
    def __init__(self, latitude: float, longitude: float):
        self.latitude = latitude
        self.longitude = longitude

    def get_tools(self) -> List[tool]:
        @tool
        def get_hourly_weather_data(from_date: date, to_date: date) -> MeteoData:
            """Get hourly weather data for a specific date range."""
            url = (f"https://api.open-meteo.com/v1/forecast?"
                   f"latitude={self.latitude}&longitude={self.longitude}&"
                   f"hourly=temperature_2m,relative_humidity_2m...")
            response = requests.get(url)
            return parse_weather_response(response.json())
        
        return [get_hourly_weather_data]

The logistics and production agents use synthetic data generators for demonstration:

class LogisticsTools:
    def get_tools(self) -> List[tool]:
        @tool
        def get_logistics_data(
            from_date: date,
            to_date: date,
            origins: Optional[List[str]] = None,
            destinations: Optional[List[str]] = None,
        ) -> LogisticsDataset:
            """Generate synthetic logistics shipment data."""
            # Generate realistic shipment data with delays, costs, routes
            records = generate_synthetic_shipments(...)
            return LogisticsDataset(records=records, aggregates=...)
        
        return [get_logistics_data]

For UI we’re going to use Chainlit. The Chainlit integration provides real-time visibility into agent execution:

class LoggingHooks(HookProvider):
    async def before_tool(self, event: BeforeToolCallEvent) -> None:
        step = cl.Step(name=f"{event.tool_use['name']}", type="tool")
        await step.send()
        cl.user_session.set(f"step_{event.tool_use['name']}", step)

    async def after_tool(self, event: AfterToolCallEvent) -> None:
        step = cl.user_session.get(f"step_{event.tool_use['name']}")
        if step:
            await step.update()

@cl.on_message
async def handle_message(message: cl.Message):
    agent = cl.user_session.get("agent")
    message_history = cl.user_session.get("message_history")
    message_history.append({"role": "user", "content": message.content})
    
    response = await agent.run_async(message.content)
    await cl.Message(content=response).send()

This creates a transparent experience where users see:

  • Which agent is handling their request
  • What tools are being invoked
  • Real-time streaming of responses

Now we can handle a variety of user queries: For example:

User: “What was the average temperature last week?”

Flow:

  1. Orchestrator identifies weather domain
  2. Routes to weather_assistant
  3. Weather agent calls get_hourly_weather_data
  4. Analyzes and returns formatted response

Or multi-domain queries:

User: “Did weather conditions affect our shipment delays yesterday?”

Flow:

  1. Orchestrator identifies weather + logistics domains
  2. Routes to weather_assistant for climate data
  3. Routes to logistics_assistant for shipment data
  4. Synthesizes correlation analysis
  5. Returns unified insight

And complex analytics:

User: “Analyze production efficiency trends and correlate with weather and logistics performance based in yesterday’s data.”

Flow:

  1. Orchestrator coordinates all three agents
  2. Production agent retrieves manufacturing KPIs
  3. Weather agent provides environmental data
  4. Logistics agent supplies delivery metrics
  5. Orchestrator synthesizes multi-domain analysis

This architecture scales naturally in multiple dimensions. We can easily add new specialized agents without disrupting existing functionality. WE only need to create the new agent and register it as a tool with the orchestratortrator prompt with new domain description. That’s it.

The orchestrator pattern transforms multi-domain AI from a monolithic challenge into a composable architecture. Each agent focuses on what it does best, while the orchestrator provides intelligent coordination.

Full code in my github.

Building an AI Frontend with Chainlit and OAuth2 Authentication

~ Gonzalo Ayuso ~ Leave a comment

Today we’ll explore how to build a secure AI frontend using Chainlit. Chainlit is Python framework that allows us to create interactive AI applications. In this example we are going to reuse the weather tool created in a previous post. Also, we will implement OAuth2 authentication with a Nginx as a reverse proxy.

The project consists of four main components:

  1. Nginx Reverse Proxy: Handles authentication via auth_request and routes traffic
  2. Fake OAuth Server: Simple Flask app that simulates OAuth2 authentication
  3. Chainlit Application: The main chat interface with AI capabilities
  4. Strands AI Agent: Weather-focused AI assistant with custom tools

The Nginx configuration implements OAuth2 authentication using the auth_request module:

server {
    listen 8000;

    location / {
        auth_request /oauth2/auth;
        
        auth_request_set $user_jwt $upstream_http_x_user_jwt;
        add_header X-Debug-User-JWT $user_jwt always;
        
        error_page 401 = @error401;
        try_files $uri @proxy_to_app;
    }

    location = /oauth2/auth {
        internal;
        proxy_pass http://oauth2/oauth2/auth;
        proxy_pass_request_body off;
        proxy_set_header Content-Length "";
        proxy_set_header X-Original-URI $request_uri;
        proxy_set_header X-Original-Remote-Addr $remote_addr;
        proxy_set_header X-Original-Host $host;
    }

    location @proxy_to_app {
        proxy_set_header X-User-JWT $user_jwt;
        proxy_pass http://chainlit;
    }
}

Key Features:

  • Every request to / triggers an authentication check via /oauth2/auth
  • JWT token is extracted from the OAuth response and forwarded to Chainlit
  • Unauthenticated users are redirected to the OAuth sign-in page
  • The JWT token is passed to Chainlit via the X-User-JWT header

A simple Flask application simulates an OAuth2 provider for demonstration purposes. In a production environment, you would replace this with a real OAuth2 provider or implemente the whole OAuth2 flow.

@app.get(f"/oauth2/auth")
def auth():
    now = datetime.now()
    response = make_response(jsonify(dict(error='OK')), 200)
    expiration = now + JWT_EXPIRATION_TIMEDELTA
    user = 'gonzalo'
    display_name = 'Gonzalo'
    response.headers['X-User-JWT'] = str(jwt.encode(dict(
        user=user,
        display_name=display_name,
        exp=int(expiration.timestamp())
    ), SECRET, algorithm=JWT_ALGORITHM))
    logger.info("Fake OAuth authentication successful")
    return response

Chainlit processes the JWT token via a custom header authentication callback:

@cl.header_auth_callback
def header_auth_callback(headers: Dict) -> Optional[cl.User]:
    if headers.get("x-user-jwt"):
        jwt_token = headers.get("x-user-jwt")
        try:
            decoded_payload = jwt.decode(jwt_token, SECRET, algorithms=[JWT_ALGORITHM])
            return cl.User(
                identifier=decoded_payload['user'],
                display_name=decoded_payload['display_name'],
                metadata={"role": 'user', "provider": "header"})
        except jwt.ExpiredSignatureError:
            cl.logger.error("Token has expired.")
            return None
    else:
        return None

This callback:

  • Extracts the JWT from the x-user-jwt header
  • Validates the token signature and expiration
  • Creates a Chainlit User object with the decoded information
  • Handles token expiration gracefully

The application uses Strands agents with both base tools and custom weather tools:

agent = get_agent(
    system_prompt=PROMPT_GENERAL,
    base_tools=get_all_base_tools(),
    custom_tools=get_all_custom_tools()
)

Base Tools Include:

  • Calculator
  • Browser access
  • Current time
  • Batch processing
  • Think (reasoning tool)

The weather functionality is implemented using custom Strands tools that fetch meteorological data:

class WeatherTools:
    def __init__(self, latitude: float, longitude: float):
        self.latitude = latitude
        self.longitude = longitude

    def get_tools(self, tools=None) -> List[tool]:
        @tool
        def get_hourly_weather_data(from_date: date, to_date: date) -> MeteoData:
            """
            Get hourly weather data for a specific date range in my city.
            
            Returns:
                MeteoData: Object containing weather readings for temperature, 
                          humidity, precipitation, etc.
            """
            # Implementation details...

The weather tools provide:

  • Hourly weather data for specific date ranges
  • Temperature readings (actual and apparent)
  • Humidity and precipitation data
  • Surface pressure measurements
  • Evapotranspiration data

The Chainlit interface provides several starter prompts to help users interact with the weather agent:

@cl.set_starters
async def set_starters():
    return [
        cl.Starter(label="Is going to rain today?", message="Is going to rain today?"),
        cl.Starter(label="tomorrow's weather", message="What will the weather be like tomorrow?"),
        cl.Starter(label="Next 7 days weather", message="Make a weather forecast for the next 7 days."),
    ]

Chainlit also supports message history management, allowing users to see their previous interactions:

@cl.on_message
async def handle_message(message: cl.Message):
    message_history = cl.user_session.get("message_history")
    message_history.append({"role": "user", "content": message.content})
    
    msg = cl.Message(content="")
    await msg.send()
    
    app_user = cl.user_session.get("user")
    question = f"user: {app_user.display_name} Content: {message.content}"
    
    async for event in agent.stream_async(question):
        if "data" in event:
            await msg.stream_token(str(event["data"]))
        elif "message" in event:
            await msg.stream_token("\n")
            message_history.append(event["message"])
    
    await msg.update()

And that’s all. Thanks to Chainlit, we can build AI frontends and integrate them with OAuth2 authentication in a secure and efficient way. The combination of Chainlit’s interactive capabilities and Nginx’s robust authentication features provides a solid foundation for building AI applications that require user authentication.

Full code in my github account

Building ReAct AI agents with sandboxed Python code execution using AWS Bedrock and LangGraph

~ Gonzalo Ayuso ~ Leave a comment

In industrial environments, data analysis is crucial for optimizing processes, detecting anomalies, and making informed decisions. Manufacturing plants, energy systems, and industrial IoT generate massive amounts of data from sensors, machines, and control systems. Traditionally, analyzing this data requires specialized knowledge in both industrial processes and data science, creating a bottleneck for quick insights.

I’ve been exploring agentic AI frameworks lately, particularly for complex data analysis tasks. While working on industrial data problems, I realized that combining the reasoning capabilities of Large Language Models with specialized tools could create a powerful solution for industrial data analysis. This project demonstrates how to build a ReAct ( Reasoning and Acting) AI agent using LangGraph that can analyze manufacturing data, understand industrial processes, and provide actionable insights.

The goal of this project is to create an AI agent that can analyze industrial datasets (manufacturing metrics, sensor readings, process control data) and provide expert-level insights about production optimization, quality control, and process efficiency. Using LangGraph’s ReAct agent framework with AWS Bedrock, the system can execute Python code dynamically in a sandboxed environment, process large datasets, and reason about industrial contexts.

The dataset is a fake sample of industrial data with manufacturing metrics like temperature, speed, humidity, pressure, operator experience, scrap rates, and unplanned stops. In fact, I’ve generated the dataset using chatgpt

This project uses several key components:

  • LangGraph ReAct Agent: For building the multi-tool AI agent with ReAct (Reasoning and Acting) patterns that can dynamically choose tools and reason about results
  • AWS Bedrock: Claude Sonnet 4 as the underlying LLM for reasoning and code generation
  • Sandboxed Code Interpreter: Secure execution of Python code for data analysis using AWS Agent Core. One tool taken from strands-agents-tools library.
  • Industrial Domain Expertise: Specialized system prompts with knowledge of manufacturing processes, quality control, and industrial IoT

The agent has access to powerful tools:

  • Code Interpreter: Executes Python code safely in a sandboxed AWS environment using pandas, numpy, scipy, and other scientific libraries
  • Data Processing: Handles large industrial datasets with memory-efficient strategies
  • Industrial Context: Understands manufacturing processes, sensor data, and quality metrics

The system uses AWS Agent Core’s sandboxed code interpreter, which means:

  • Python code is executed in an isolated environment
  • No risk to the host system
  • Access to scientific computing libraries (pandas, numpy, scipy)
  • Memory management for large datasets

The core of the system is surprisingly simple. The ReAct agent is built using LangGraph’s create_react_agent with custom tools:

from langgraph.prebuilt import create_react_agent
from typing import List
import pandas as pd
from langchain_core.callbacks import BaseCallbackHandler


def analyze_df(df: pd.DataFrame, system_prompt: str, user_prompt: str,
               callbacks: List[BaseCallbackHandler], streaming: bool = False):
    code_interpreter_tools = CodeInterpreter()
    tools = code_interpreter_tools.get_tools()

    agent = create_react_agent(
        model=get_llm(model=DEFAULT_MODEL, streaming=streaming,
                      budget_tokens=12288, callbacks=callbacks),
        tools=tools,
        prompt=system_prompt
    )

    agent_prompt = f"""
    I have a DataFrame with the following data:
    - Columns: {list(df.columns)}
    - Shape: {df.shape}
    - data: {df}
    
    The output must be an executive summary with the key points.
    The response must be only markdown, not plots.
    """
    messages = [
        ("user", agent_prompt),
        ("user", user_prompt)
    ]
    agent_input = {"messages": messages}
    return agent. Invoke(agent_input)

The ReAct pattern (Reasoning and Acting) allows the agent to:

  1. Reason about what analysis is needed
  2. Act by calling the appropriate tools (in this case: code interpreter)
  3. Observe the results of code execution
  4. Re-reason and potentially call more tools if needed

This creates a dynamic loop where the agent can iteratively analyze data, examine results, and refine its approach – much more powerful than a single code execution.

The magic happens in the system prompt, which provides the agent with industrial domain expertise:

SYSTEM_PROMPT = """
# Industrial Data Analysis Agent - System Prompt

You are an expert AI agent specialized in industrial data analysis and programming. 
You excel at solving complex data problems in manufacturing, process control, 
energy systems, and industrial IoT environments.

## Core Capabilities
- Execute Python code using pandas, numpy, scipy
- Handle large datasets with chunking strategies  
- Process time-series data, sensor readings, production metrics
- Perform statistical analysis, anomaly detection, predictive modeling

## Industrial Domain Expertise
- Manufacturing processes and production optimization
- Process control systems (PID controllers, SCADA, DCS)
- Industrial IoT sensor data and telemetry
- Quality control and Six Sigma methodologies
- Energy consumption analysis and optimization
- Predictive maintenance and failure analysis
"""

The code interpreter tool is wrapped with safety validations:

def validate_code_ast(code: str) -> bool:
    """Validate Python code using AST to ensure safety."""
    try:
        ast.parse(code)
        return True
    except SyntaxError:
        return False


@tool
def code_interpreter(code: str) -> str:
    """Executes Python code in a sandboxed environment."""
    if not validate_code_ast(code):
        raise UnsafeCodeError("Unsafe code or syntax errors.")

    return code_tool(code_interpreter_input={
        "action": {
            "type": "executeCode",
            "session_name": session_name,
            "code": code,
            "language": "python"
        }
    })

The system uses Claude Sonnet 4 through AWS Bedrock with optimized parameters for industrial analysis:
def get_llm(model: str = DEFAULT_MODEL, max_tokens: int = 4096,
            temperature: float = TemperatureLevel.BALANCED,
            top_k: int = TopKLevel.DIVERSE,
            top_p: float = TopPLevel.CREATIVE) -> BaseChatModel:
    model_kwargs = {
        "max_tokens": max_tokens,
        "temperature": temperature,
        "top_k": top_k,
        "top_p": top_p
    }

    return ChatBedrock(
        model=model,
        client=aws_get_service('bedrock-runtime'),
        model_kwargs=model_kwargs
    )

The project includes fake sample industrial data with manufacturing metrics:

- `machine_id`: Equipment identifier
- `shift`: Production shift (A/M/N for morning/afternoon/night)
- `temperature`, `speed`, `humidity`, `pressure`: Process parameters
- `operator_experience`: Years of operator experience
- `scrap_kg`: Quality metric (waste produced)
- `unplanned_stop`: Equipment failure indicator

A typical analysis query might be: "Do temperature and speed setpoints vary across shifts?"
The agent will stream the response as it generates it.

The agent will:

1. Load and examine the dataset structure
2. Generate appropriate Python code for analysis
3. Execute the code in a sandboxed environment
4. Provide insights about shift-based variations
5. Suggest process optimization recommendations
import logging

import pandas as pd
from langchain_core.callbacks import StreamingStdOutCallbackHandler

from modules.df_analyzer import analyze_df
from prompts import SYSTEM_PROMPT

logging.basicConfig(
    format='%(asctime)s [%(levelname)s] %(message)s',
    level='INFO',
    datefmt='%d/%m/%Y %X')

logger = logging.getLogger(__name__)


class StreamingCallbackHandler(StreamingStdOutCallbackHandler):
    def on_llm_new_token(self, token: str, **kwargs):
        print(token, end='', flush=True)


df = pd.read_csv('fake_data.csv')

user_prompt = "Do temperature and speed setpoints vary across shifts?"
for chunk in analyze_df(
        user_prompt=user_prompt,
        df=df,
        system_prompt=SYSTEM_PROMPT,
        callbacks=[StreamingCallbackHandler()],
        streaming=True):
    logger.debug(chunk)

This project demonstrates the power of agentic AI for specialized domains. Instead of building custom analytics dashboards or writing specific analysis scripts, we provide the agent with:

  1. Domain Knowledge: Through specialized system prompts
  2. Tools: Safe code execution capabilities
  3. Context: The actual data to analyze

The agent can then:

  • Generate appropriate analysis code
  • Execute it safely
  • Interpret results with industrial context
  • Provide actionable recommendations

The result is a flexible system that can handle various industrial analysis tasks without pre-programmed solutions. The agent reasons about the problem, writes the necessary code (sandboxed), and provides expert-level insights.

Full code in my github.

Agentic AI for movie recommendations with Python and Strands Agents

~ Gonzalo Ayuso ~ Leave a comment

Context 1: I like to go to the cinema. I normally go to the cinema on Saturday afternoons, at the first showing. In the city where I live there are three cinemas and all belong to the same company called Sade. I normally check the cinema schedules on their website, SadeCines.com, to see what’s playing. Also, I track the movies I see on Letterboxd. There I have my diary and also a list with the movies I see in the cinema. I rate the movies when I finish watching them. My first impression. I do that not to share with others, only to have a personal record of what I like and dislike.

Context 2: I’m on holidays and I like to code also, so I decided to build an AI agent that helps me decide what movie to watch on Saturday afternoons. This project is an example of over-engineering, I know, but I’ve done it as an exercise using Strands Agents, a framework for building multi-tool LLM agents that I’m using these days.

The aim of the project is to create an AI agent that can access the internet to check the cinema schedules, my Letterboxd profile, and then recommend me a movie to watch on Saturday afternoons. Normally the LLMs are good at reasoning, but they don’t have access to the internet. Also, they are not good at doing mathematical operations, but with agents we can use tools to do that. So I decided to build an agent that can access the internet (to check the cinema schedules, my Letterboxd profile and IMDb/Metacritic’s scores) and create the needed code to do the mathematical operations needed.

Strands Agents (it is similar to LangChain) allows us to build multi-tool LLM agents. In this example I’m using the pre-built tools provided by the framework, like:

  • calculator: for performing mathematical operations
  • think: for reasoning and decision-making
  • current_time: to get the current date and time
  • file_write: to write the recommendations to a file
  • batch: to execute multiple tools in parallel
  • code_interpreter: to execute Python code dynamically (sandboxed in an AWS environment)
  • browser: to scrape the cinema schedules from SadeCines.com and my Letterboxd profile (also sandboxed in an AWS environment)

Code interpreter is a powerful tool that allows us to execute Python code dynamically, which is useful for performing mathematical operations and data processing. For me it is the key to push the agents to the next level. LLMs can generate python code very well. They can generate code to build a Pandas dataframe, to filter the data, to calculate the average rating, etc. But they can also generate code that can be harmful, like deleting files, or accessing sensitive data. So we need to be careful with the code we execute. This issue is especially important when we are using prompts from users (in a chat, for example). Strands Agents provides a tool called python-repl that allows us to execute Python code locally within our environment. If you rely on your prompts it can be an option (I’ve sent a pull request to Strands Agents to make it a bit more safe). But in this project I’m using the code_interpreter tool, which is a sandboxed environment provided by AWS. This allows us to execute Python code safely without the risk of executing harmful code in your host environment.

In this project we need to scrape webpages to retrieve information from internet. Strands Agents provides us a built-in tool, called use_browser, to use a headless browser locally to access the Internet. In this project, I’m using the browser tool, which is also a sandboxed environment provided by AWS Bedrock. This allows us to scrape webs (using Playwright) and interact with web pages without the risk of executing harmful code in your host environment.

With this information, to build the agent is pretty straightforward. The idea of agents is not to code everything from scratch, but to provide to the agent the needed tools to solve the problem, and let the agent figure out how to use them using the prompts. When we work with LLM we have two kinds of prompts: the system prompt and the user prompt. The system prompt is used to define the agent’s behavior, while the user prompt is used to provide the input data.

In this project I’m using those prompts:

from settings import BASE_DIR

SYSTEM_PROMPT = f"""
You are an expert movie recommendation assistant to help me decide what to watch.

You have access to the following URLs and available movie analyses:
- https://sadecines.com/ With the movie schedules in my city's cinemas.
    Sadecines has a checkbox to filter the day of the week, so you can select Saturday.
- https://letterboxd.com/gonzalo123/films/diary/ Movies I have watched and rated.
- https://letterboxd.com/gonzalo123/list/cine-2025/detail/ Movies I have already seen in theaters in 2025.

You must take into account the user's preferences:
- Avoid movies in the "children" and "family" genres.
- I don't really like intimate or drama movies, except for rare exceptions.
- I like entertaining movies, action, science fiction, adventure, and comedies.

Take into account when making recommendations:
- The ratings of the movies on IMDb and Metacritic.
- But mainly consider my personal preferences,
    which can be seen in the list of movies I have watched and rated on Letterboxd.
"""

QUESTION = f"""
Analyze the movies showing this Saturday in the first session.

Present only those you recommend, excluding those not relevant according to my preferences,
and order them from best to worst according to your criteria.

Show the result in a table with the following columns:
- Title
- Genre
- IMDb Rating
- Metacritic Rating
- Summary
- Start Time
- End Time

Save the final report in a file named YYYYMMDD.md, following this structure:
{BASE_DIR}/
    └ reports/
        └ YYYYMMDD.md       # Movie analysis of the day, format `YYYYMMDD`

And the code of the agent is very simple (I’m using AWS Bedrock to run the agent)

import logging

from botocore.config import Config
from strands import Agent
from strands.models import BedrockModel
from strands_tools import calculator, current_time, think, file_write, batch
from strands_tools.browser import AgentCoreBrowser
from strands_tools.code_interpreter import AgentCoreCodeInterpreter

from promts import SYSTEM_PROMPT, QUESTION
from settings import AWS_REGION, MODEL_TEMPERATURE, MODEL, LLM_READ_TIMEOUT, LLM_CONNECT_TIMEOUT, LLM_MAX_ATTEMPTS

logging.basicConfig(
    format="%(asctime)s [%(levelname)s] %(message)s",
    level="INFO",
    datefmt="%d/%m/%Y %X",
)

logger = logging.getLogger(__name__)

agent = Agent(
    system_prompt=SYSTEM_PROMPT,
    model=BedrockModel(
        model_id=MODEL,
        temperature=MODEL_TEMPERATURE,
        boto_client_config=Config(
            read_timeout=LLM_READ_TIMEOUT,
            connect_timeout=LLM_CONNECT_TIMEOUT,
            retries={'max_attempts': LLM_MAX_ATTEMPTS}
        )
    ),
    tools=[
        calculator, think, current_time, file_write, batch,
        AgentCoreCodeInterpreter(region=AWS_REGION).code_interpreter,
        AgentCoreBrowser(region=AWS_REGION).browser]
)

result = agent(QUESTION)
logger.info(f"Total tokens: {result.metrics.accumulated_usage['totalTokens']}")
logger.info(f"Execution time: {sum(result.metrics.cycle_durations):.2f} seconds")
logger.info(f"Tools used: {list(result.metrics.tool_metrics.keys())}")

The lines of code never is a goal (we only need to write readable and maintainable code), but in this example we have more code in the prompts than in the code itself. Maybe it’s the sigh of our times.

And that’s all. I must say again that this project is just an example. It is an over-engineering example. Scaling this project would be very expensive. Working a little bit in a custom scraper in addition to custom python code, can do the same to solve this specific problem without the usage, and paid, the IA (cheap for a single user usage, but expensive when scaled). I think it is a good example to show how Agents and the power of the code interpreter and the browser tools in a few lines of code. And remember, I’m on holidays and I like to code (don’t blame me for that).

Full code in my Github account.