DEV Community

Cover image for End to End Generative AI mini Project On AWS Using AWS Bedrock, AWS Lambda, API Gateway and S3
AWS Community Builders profile image Tapas SInghal
Tapas SInghal for AWS Community Builders

Posted on

End to End Generative AI mini Project On AWS Using AWS Bedrock, AWS Lambda, API Gateway and S3

#aws #bedrock #ai #serverless

In this blog, I’ll walk you through exactly how I created an AI-powered blog generator bot using Amazon Bedrock, Lambda, S3, and API Gateway. This was a mini project to practice serverless architecture using AWS services and AI text generation models to get introduce to Amazon Bedrock.

🔧 I used a non-chat LLaMA model from Amazon Bedrock, because that was available in my region (ap-south-1) and suited my use case well for blog generation, dont hesitate to use anyother model.

✨ What You’ll Learn

  • How to access and use Amazon Bedrock models
  • How to prepare a Lambda function with external libraries like boto3
  • How to integrate Lambda with API Gateway for POST requests
  • How to store generated blog content into S3 buckets
  • How to test the endpoint using free tools like ReqBin

🛠️ Step 1: Enable Amazon Bedrock and Choose Model

🔹 Accessing Bedrock

  1. Go to AWS Console → Services → Amazon Bedrock.
  2. Request model access (you may need to request access for models like LLaMA3, Titan, etc.).
  3. Wait for approval (this can take from minutes to a few hours).

💡 I used meta.llama3-8b-instruct-v1:0, which was available in my region. I didn’t choose a chat model specifically — I used what was available for experimentation. If available you should choose a chat based model or any other good text based model.

Amazon Bedrock Model Access

🧠 Step 2: Write and Prepare the Lambda Function

🔹 Writing the Function

The Lambda function handles:

  • Receiving the blog topic from API Gateway
  • Sending the prompt to the Bedrock model
  • Receiving and parsing the response
  • Saving the blog to S3 
import boto3
import botocore.config
import json
from datetime import datetime
def blog_generate_using_bedrock(blogtopic:str)->str:
    prompt=f"""Write a blog minimum 300 to 500 words(no need to strictly follow) in a professional mammer with only 5 subtopics and a conclusion {blogtopic}"""
    body={
        "prompt":prompt,
        "max_gen_len":600,
        "temperature":0.5,
        "top_p":0.9
    }
    try:
        bedrock=boto3.client("bedrock-runtime",region_name="ap-south-1",config=botocore.config.Config(read_timeout=300,retries={'max_attempts':3}))
        response=bedrock.invoke_model(body=json.dumps(body),modelId="<Your_Model_ID>")

        response_content=response.get('body').read()
        response_data=json.loads(response_content)
        print(response_data)
        blog_details=response_data['generation']
        return blog_details
    except Exception as e:
        print(f"Error generating the blog:{e}")
        return ""

def save_blog_details_s3(s3_key,s3_bucket,generate_blog):
    s3=boto3.client('s3')
    try:
        s3.put_object(Bucket = s3_bucket, Key = s3_key, Body =generate_blog )
        print("Code saved to s3")

    except Exception as e:
        print("Error when saving the code to s3")

def lambda_handler(event, context):
    # TODO implement

    event=json.loads(event['body'])
    blogtopic=event['blog_topic']
    generate_blog=blog_generate_using_bedrock(blogtopic=blogtopic)
    if generate_blog:
        curr_time=datetime.now().strftime('%H%M%S')
        s3_key=f"blog-output/{curr_time}.txt"
        s3_bucket='<Your_s3_bucket>'
        save_blog_details_s3(s3_key,s3_bucket,generate_blog)
    else:
        print("No blog was generated ")

    return{
        'statusCode':200,
        'body':json.dumps('Blog Generation is Completed')
    }
Enter fullscreen mode Exit fullscreen mode

NOTE: - Replace 'Your_Model_ID' with your bedrock model's ModelID and 'Your_s3_bucket' with your AWS bucket

🔹 Important: Add boto3 Dependency

Amazon Bedrock needs latest boto3 version, which is not included in Lambda by default. You have two options:

✅ Option 1: Create a Lambda Layer with boto3

  1. On your local system, run: 
   mkdir python
   pip install boto3 -t python/
   zip -r boto3_layer.zip python
Enter fullscreen mode Exit fullscreen mode
  1. Go to AWS Console → Lambda → Layers → Create Layer
  2. Upload boto3_layer.zip and choose compatible runtimes (Python 3.12, 3.11, or 3.10)
  3. Attach the layer to your Lambda function

Your Result should look like this
layer

🧠 Option 2: Include boto3 manually in deployment zip

(Same as above, but upload the entire zipped folder as your Lambda deployment package.)

🔗 Step 3: Setup API Gateway for Integration

🔹 Create an HTTP API

  1. Go to AWS Console → API Gateway
  2. Choose HTTP API
  3. Click “Build” and name it (e.g., BlogGenAPI)

🔹 Configure Routing

  1. Create a route: POST /blog-generation
  2. Add integration target: Your Lambda function
  3. Deploy the API and create a Stage (in my case: dev)
  4. You’ll get 
   Invoke URL
   https://your-api-id.execute-api.region.amazonaws.com/dev/

   AND 

   Route details
   POST /blog-generation
Enter fullscreen mode Exit fullscreen mode

Stage Deploy

API Route integration with lambda

🪣 Step 4: Create and Connect S3 Bucket

  1. Go to AWS Console → S3 → Create a bucket
    • Bucket name: awsbedrocktapas (you can name it anything but should be unique)
  2. No need to make it public
  3. Go to IAM role of the Lambda and allow:
    • s3:PutObject on your bucket

s3 Bucket

In your Lambda code, the S3 upload part looks like this:

s3_key = f"blog-output/{curr_time}.txt"
s3_bucket = 'YOUR Bucket'
s3.put_object(Bucket=s3_bucket, Key=s3_key, Body=generated_blog)
Enter fullscreen mode Exit fullscreen mode

Raplace 'YOUR Bucket' with your bucket name you just created

🧪 Step 5: Test the API using any tool, I am using ReqBin (Free Online Tool)

  1. Go to reqbin
  2. Choose method: POST
  3. URL: Your deployed API Gateway endpoint 
https://your-api-id.execute-api.region.amazonaws.com/dev/blog-generation
Enter fullscreen mode Exit fullscreen mode
  1. Choose Content-Type: json
  2. Add body: 
{
  "blog_topic": "Benefits of Remote Work in 2025"
}
Enter fullscreen mode Exit fullscreen mode
  1. Hit “Send” and check the response

API Testing

You should see:

{
  "blog_topic": "Benefits of Remote Work in 2025"
}
Enter fullscreen mode Exit fullscreen mode

Now check your S3 bucket — you’ll find a .txt file with your freshly AI-generated blog!

S3 output files

🧾 Example Output Blog

Blog Output

✅ Final Flow Diagram (Optional)

Architecture

🎯 Conclusion

Using just four core AWS services — Bedrock, Lambda, API Gateway, and S3 — we’ve built a complete, serverless AI-powered blog generator.

This project taught me how to:

  • Work with LLMs via Bedrock
  • Integrate custom code into Lambda
  • Manage external Python libraries via Lambda Layers
  • Expose a POST API endpoint for external use
  • Save data persistently using S3

💬 You can build on top of this by adding:

  • A frontend (React/Next.js)
  • An email sender to notify when blog is ready
  • An admin dashboard for blog management

🌐 My Socials

Feel free to connect with me or collaborate on other serverless & AI projects:

📂 Project Repository

🔗 GitHub: https://github.com/its-tapas/blogGen

🤝 Open to improvements, ideas, and suggestions. Don't hesitate to fork, star, or contribute to this project!

Thanks for reading 🙌 — Tapas Singhal

Top comments (0)