Inspiration

AutoSpec.AI is a serverless application that enables users to email or upload documents (PDF, DOCX, TXT) and receive structured, AI-generated system requirements using a Bedrock-powered prompt that mimics the mindset of a skilled systems analyst. The output includes functional requirements, non-functional requirements, stakeholder roles, and technical transparency in both narrative and tabular format.

What it does

  1. Receive document input (via email or API upload).
  2. Store document in S3.
  3. Trigger Lambda function on new upload.
  4. Parse document (PDF, DOCX, TXT).
  5. Call Amazon Bedrock with a well-crafted prompt.
  6. Format AI response into structured system requirements (Markdown, JSON, optionally PDF).
  7. Send result back via email (SES) or Slack webhook.
  8. Save input/output history in DynamoDB with metadata for version tracking. ## How we built it Used Amazon Q and Claude Code to build the deep AWS integrations.

Challenges we ran into

Accomplishments that we're proud of

The full Lambda architecture that leverages other AWS services such as S3 and DyanamoDB.

What we learned

Leave more time for the project publishing.

What's next for Auto Spec AI

Moving into full production mode and marketing it.

Share this project:

Updates