Proof Butler

• 

  Home screen of Proof Butler showing the three core actions: create a proof, verify a proof or view existing proofs.

• 

  Creating a proof by selecting an image and assigning a title. The app generates a SHA-256 hash and stores it as a tamper-check record.

• 

  List of saved proofs with title, thumbnail, creation time, and a shortened hash. Each proof can be copied and shared using its Proof ID.

• 

  Verification by selecting a Proof ID and comparing it with a chosen image. A mismatch indicates the file was altered.

About the project

The problem

During a hackathon, I kept thinking about how easy it is to lose trust in an image once it starts getting shared. A photo gets forwarded, screenshot, edited, cropped, or re saved, and after a while nobody can confidently say which version is the original. Even if you trust the person, you cannot trust the file.

I wanted a tiny, practical tool that does one thing well which is creating a stable fingerprint for an image, then let you check any later copy against that fingerprint.

What Proof Butler does

Proof Butler is a Flutter app powered by a Serverpod backend that creates and verifies tamper checks for images.

It has three flows:

Create Proof

• Pick an image from your device
• Enter a required title (so you can recognize it later)
• The app computes a SHA-256 hash locally from the image bytes
• The app generates a small thumbnail preview
• It saves a proof record to the backend:
  • title
  • SHA-256 hash
  • createdAt timestamp
  • thumbnail bytes
• You can copy the Proof ID and details for sharing

My Proofs

• Lists saved proofs with thumbnail + title
• Shows ID, createdAt, and a short hash preview
• One tap copy to share proof details

Verify Proof

• Select a stored proof record
• Pick an image file you want to check
• The app computes SHA-256 for that file and asks the backend to compare it
• Result is clear:
  • MATCH (same bytes)
  • MISMATCH (file differs)

This is not an image similarity detector but a strict integrity check. If the bytes changed even slightly, the hash changes.

How I built it

• Flutter (Material 3) for the UI
• file_picker to pick images
• crypto to compute SHA-256 on device
• image package to decode and create a compact JPEG thumbnail
• Serverpod backend with simple endpoints:
  • createEvidenceRecord (save title, hash, thumbnail)
  • listEvidenceRecords (load proofs)
  • verifyEvidence (compare stored hash with computed hash)
• PostgreSQL via Serverpod for persistence

The phone computes the hash and the server stores and compares.

Challenges

• Async mistakes early on (missing await caused UI to update before the server response came back)
• Handling bytes cleanly across PlatformFile, Uint8List, and ByteData

What I am proud of

• End to end workflow that is easy to demo: create, list, verify match, verify mismatch
• Titles + thumbnails make old records recognizable without storing the original image
• Serverpod made it fast to go from models to a working typed client

What’s next

• Verify by Proof ID (paste or scan) so someone can verify without browsing your list
• Share links or QR codes for Proof IDs
• User accounts or access controls if this becomes multi user
• Optional metadata like tags or capture context

Built With

• dart
• docker
• flutter
• postgresql
• serverpod