Launching today
ThreadLore
Turn fabric photos into editable loom drafts
37 followers
Turn fabric photos into editable loom drafts
37 followers
When a textile’s weaving instructions are lost, a photograph still holds clues. ThreadLore helps handweavers inspect a fabric photo, check a suggested weave against the yarn crossings, and turn the pattern they approve into a printable loom card or WIF file.




Launch Team

AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
Hi everyone! 👋
I built ThreadLore for handweavers and craft enthusiasts who have a fabric sample or photograph but no longer have the original loom instructions.
Reconstructing a weave manually from a photo takes hours of squinting at individual threads. ThreadLore automates the heavy lifting:
1. OpenCV on Google Cloud Run detects thread spacing and extracts yarn colors.
2. Gemini Flash on Vertex AI suggests the repeating over-under pattern.
3. The weaver can inspect the overlay and click individual crossings to correct any errors.
4. A deterministic compiler calculates the four-shaft loom setup, mathematically checks every crossing, and exports a printable loom card and a WIF 1.1 file for weaving software.
We built this for the Google Cloud Run "Build for Good" hackathon.
Try it live: https://threadlore-122120455814.us-central1.run.app/
Explore the code: https://github.com/KenEverestNunes/ThreadLore
Demo video: https://youtu.be/QQ4wO8BPFlU
I'd love to hear your thoughts and feedback!
being able to check the suggested weave against the actual yarn crossings matters more than getting a convincing looking pattern
the wif export gives that check somewhere useful to go
How many weavers did you watch rebuild a draft by hand before you built this? Curious which part they hated more, counting the threads or checking the crossings.
the deterministic check after the model guesses is the right call. we do the same for 3d prints, ai suggests, then a boring check makes sure it actually prints