Skip to content

Lucineer/fishinglog-ai

Repository files navigation

fishinglog-ai 🎣

You stop writing good catch notes after the third beer. This remembers for you. Upload a photo for a species ID or ask a question about local techniques. It runs on your infrastructure and commits every interaction directly to your own repository.

Live Instance: fishinglog-ai.casey-digennaro.workers.dev

Fork this repository, deploy it to a Cloudflare Worker with your API keys, and start logging. Every conversation and correction is saved as a commit to your fork. No one else holds your data.

Why This Exists

Most fishing apps monetize your log data. This one doesn’t. It was built for anglers who prefer trusting their GitHub account over a third-party platform.

Quick Start

  1. Fork this repository to your GitHub account.
  2. Clone your fork locally:
    git clone https://github.com/your-username/fishinglog-ai
    cd fishinglog-ai
    npm install
  3. Log in to Cloudflare:
    npx wrangler login
  4. Set your required secrets (you only do this once):
    # A GitHub token with 'repo' permissions
    npx wrangler secret put GITHUB_TOKEN
    # Your DeepSeek or other LLM API key
    npx wrangler secret put DEEPSEEK_API_KEY
  5. Deploy the Worker:
    npx wrangler deploy

Your instance is live in about 10 seconds. It only communicates with your repo and your chosen LLM API.

What It Does

  • Identify Fish: Provides species identification for common North American saltwater and freshwater game fish from your photos.
  • Answer Questions: Responds to queries about techniques, regulations, or conditions based on the notes you’ve previously logged.
  • Voice Logging: Dictate catch notes hands-free.
  • Own Your Data: All logs are saved as Markdown files in your repository. There is no separate database or export step.
  • Swap Models: Configure it to use different compatible LLM APIs.

Honest Limitation

Species identification is optimized for common North American game fish. It may struggle with rare species, juveniles, or photos where the fish occupies less than 10% of the frame.

Architecture

This is a stateless agent built on the Cocapn Fleet protocol, running on a Cloudflare Worker. It has zero runtime dependencies and is MIT licensed.

About

fishinglog.ai — Edge AI fishing vessel. Jetson-powered species classification, captain voice, conversational training. Powered by cocapn.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors