PitchPulse

Forecast smarter. Pitch better. Know your value.

Comment

Inspiration

As an aspiring founder, I constantly faced a familiar challenge: “How much is my startup worth, and how do I present it professionally?”

Valuation calculators are either too vague or too complex. Pitch decks take time to design. And no one tells you when you’ll run out of cash. We set out to build an AI-powered Startup Copilot that can:

  1. Estimate your startup’s valuation

  2. Auto-generate investor-ready decks

  3. Show a visual financial forecast based on your burn and runway

What it does

PitchPulse helps founders go from raw business metrics to investor-ready assets in minutes.

  1. Startup Valuation Takes in basic inputs: monthly revenue, market type, and growth rate. Uses ARR projection and revenue multiples to estimate startup valuation. Powered by Claude (via Amazon Bedrock) for explainable, realistic output. Saves data to DynamoDB and optionally logs events to EventBridge.

  2. Pitch Deck Generator Automatically creates professional pitch decks in seconds. Supports multiple formats: Elevator Pitch, Investor Pitch, Stakeholder Meeting Deck, Founding Team Introduction User can choose style: slide format or bullet points.

    1. Financial Forecastor Visualizes your financial future with a dynamic chart: Monthly profit/loss line, Cash runway tracker, Break-even point marker, Burn rate visualization Accepts inputs like starting cash, burn rate, and monthly growth. Chart is generated as an image and uploaded to S3, with a public link returned.
    2. Built-in Summarization Any valuation output can be instantly summarized into concise bullet points. Helpful for quick review, investor teasers, or internal use.

How we built it

This project was built using a serverless architecture on AWS, combining frontend design, backend orchestration, AI generation, and data visualization:

Backend AWS Lambda Core logic lives in modular Lambda functions: TriggerValuationWorkflow: Handles frontend requests and starts a Step Function. ValidateValuationInput: Ensures inputs are present and valid. GenerateValuation: Uses Amazon Bedrock (Claude) to generate the valuation response. SaveValuationData: Stores valuation in DynamoDB and emits EventBridge events. GeneratePitchDeck: Dynamically creates investor/elevator/stakeholder pitch decks using Claude. SummarizeValuation: Summarizes valuation into 4–5 bullet points. GenerateForecastGraph: Accepts financial data and returns a base64 chart, which is uploaded to S3.

AWS Step Functions (Express)

Orchestrates the valuation pipeline in 3 steps: Input validation Claude-based valuation generation DynamoDB saving and optional event triggers

Amazon Bedrock (Claude) Powers all natural language generation tasks: Valuation report, Pitch decks, Valuation summary (bullet points), Amazon DynamoDB, securely stores startup data and generated valuations.

Amazon EventBridge Publishes ValuationCompleted events for downstream consumers (like analytics, alerts).

Amazon S3 Stores financial forecast graph images and serves them via public URLs.

Frontend Next.js + Tailwind-style design

AWS Amplify: Frontend is deployed on AWS Amplify for fast, scalable, and CI-integrated hosting. Enables automatic deployment from GitHub pushes.

API Gateway (HTTP API): Routes requests from frontend to Lambda-backed endpoints securely, with CORS enabled.

Challenges we ran into

  1. Step Function Payload Design
  2. Handling Bedrock Responses
  3. Standard vs. Express Step Functions
  4. Base64 Graph Upload to S3
  5. Custom Lambda Layers for Graph Libraries
  6. Multi-Lambda Coordination

Accomplishments that we're proud of

  1. End-to-End Serverless Architecture: We successfully built a fully serverless system using AWS Lambda, Step Functions, API Gateway, S3, DynamoDB, and EventBridge—all working in harmony without managing any servers.
  2. Intelligent AI-Driven Valuation: We used Amazon Bedrock to power our valuation logic through Claude, allowing users to receive customized startup valuations based on inputs like revenue, market, and growth.
  3. ** Auto-Generated Investor-Ready Pitch Decks**
  4. Graphical Financial Forecastor
  5. Deployed & Scalable Frontend

What we learned

We gained deep insights into building serverless architectures using AWS. We learned how to use Lambda functions to handle logic, connect services, and ensure scalable execution without managing servers. We understood the role of IAM policies in securely invoking Bedrock models and accessing other AWS services. Additionally, we explored easy deployment strategies using tools like Amplify for frontend and focused on clean, modular API design with minimal infrastructure overhead.

What's next for PitchPulse

We plan to expand PitchPulse into a dual-sided platform for both founders and investors. Upcoming features include investor-matching, downloadable pitch decks, financial forecast insights, and AI-generated due diligence reports. This will make PitchPulse a one-stop solution for early-stage fundraising and investment analysis.

Built With

  • apigateway
  • aws-lambda
  • awsamplify
  • claude
  • dynamodb
  • eventbridge
  • github
  • next.js
  • python
  • s3
  • step-functions
Share this project:

Updates