Project TechRecon designed to help enterprise leaders master the overwhelming pace of technological change. Addressing a packed executive audience at re:invent 2025, we're introducing a methodology that reframes technology scouting from a manual, reactive task into a dynamic, automated strategic intelligence operation. You can create Part1: Landscape Analysis, Part2: Technical Position Papers.
- Presented at re:Invent 2025
- A leader's guide to emerging technologies: From insights to rapid action(Track SNR203)
| Part 1: Landscape Analysis: Comprehensive scan of Emerging Technologies for your business |
|---|
![]() |
| Part 2: Technical Position Papers |
|---|
![]() |
A web application to run and monitor the Tech Recon project.
- Real-time Log Streaming: Monitor execution progress in real-time through WebSocket
- Part1/Part2 Execution: Run processes with button clicks in the web UI
- File Download: Download generated artifact files individually or as ZIP archives
- Visual Monitoring: View execution status, log count, file count at a glance
# Install uv (macOS/Linux)
curl -LsSf https://astral.sh/uv/install.sh | sh
# Or use Homebrew (macOS)
brew install uv
# Verify installation
uv --version# Navigate to project directory
cd tech-recon
# Create virtual environment and install dependencies (based on pyproject.toml)
uv syncThis single command automatically installs all web-related packages including Flask.
Create a .env file in the project root directory and configure the following environment variables:
# Create .env file
cd aws-samples/sample-tech-recon/tech-recon
touch .env
cat > /tech-recon/.env << 'EOF'
COMPANY_NAME=Pfizer ### Edit here to add your company name
INDUSTRY=pharmaceutical/healthcare ### Edit here to add your industry
TAVILY_API_KEY= ### Paste your Tavily API Key
EOFAWS Configuration (Required)
# AWS Region
AWS_DEFAULT_REGION=us-west-2
# AWS Profile (optional, not needed when using IAM Role)
AWS_PROFILE=your-profile-name
This project supports the following AWS authentication methods:
-
IAM Role (Recommended for SageMaker environment)
- Set only
AWS_DEFAULT_REGIONin.envfile - IAM Role automatically handles authentication
- Set only
-
AWS Profile
- Configure profile in
~/.aws/credentials - Set
AWS_PROFILEin.envfile
- Configure profile in
-
Environment Variables
- Directly set
AWS_ACCESS_KEY_IDandAWS_SECRET_ACCESS_KEY(not recommended)
- Directly set
Your AWS account/role must have the following permissions:
bedrock:InvokeModelbedrock:InvokeModelWithResponseStream- Other Bedrock-related read permissions
# Navigate to project directory
cd aws-samples/sample-tech-recon/tech-recon
# Run the start script
./start_web.shWhen the web server starts, you'll see this message:
============================================================
Tech Recon Web Application
============================================================
Server starting at: http://localhost:5000
Press Ctrl+C to stop the server
============================================================
# Navigate to project directory
cd aws-samples/sample-tech-recon
# Activate virtual environment
source uv run python main.pyIf you run the code in terminal, all files will be saved in artifacts/part1 or artifacts/part2.
http://localhost:5000
Since it's running locally, access the above address in your browser.
- Click Run Part1 or Run Part2(only after the part1) button
- Click "OK" in the confirmation dialog
- Monitor progress in the real-time log screen
- Real-time Logs: View logs in real-time in the left panel
- Auto-scroll: Enabled by default, can be toggled
- Clear Logs: Clear the screen with the "Clear Logs" button
- Click the "Download" button for each file in the "Generated Files" section on the right panel
- Download Part1: Download all files generated in Part1 as ZIP
- Download Part2: Download all files generated in Part2 as ZIP
- Download All: Download all files from Part1 and Part2 in a single ZIP
- Firewall: Port 5000 must be open
- Concurrent Execution: Only one Part can run at a time (Part1 -> Part2)
- Browser Compatibility: Latest versions of Chrome, Firefox, Safari, Edge recommended
- Network: Accessible only from local network (0.0.0.0:5000)
This project follows the license of the original Tech Recon project.
If you encounter any issues or have suggestions for improvements, please submit an issue.
This project is originally developed from Deep Insight Korea SA team.


