clippy-vision
https://github.com/protocorn/clippy-vision
Python
Fully local AI assistant that watches your screen to eliminate the context problem. No cloud. No data leakage.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported1 Subscribers
Add a CodeTriage badge to clippy-vision
Help out
- Issues
- Ollama calls don't set num_ctx, so long prompts may be silently truncated
- fix: send explicit Ollama num_ctx on every generate/chat call
- Linux AppImage: old mcp-runtime-<hash> folders are never cleaned up, and clippy-mcp is rewritten on every launch
- Linux AppImage: MiniLM embedding model is looked up inside the code folder, so semantic search re-downloads it or silently uses mismatched hash vectors
- Ollama address from Settings never reaches the Python API or MCP server (always uses 127.0.0.1:11434)
- First MiniLM download pulls ~932 MB (every format) when ~91 MB is needed
- fix(embeddings): download only MiniLM safetensors and tokenizer files
- Settings: reject an empty Ollama address instead of silently resetting to default
- "Ollama server is not running" error should name the address it tried
- First request after an Ollama address change fails during the ~4.6 s API restart
- Docs
- Python not yet supported