Hi HN,
I built Ch to bring AI directly into the terminal, offering fast, lightweight functionality that remains fully under user control. Most AI tools feel detached from developer workflows, but Ch is designed to integrate seamlessly, with no hidden automation or costs.
Vision:
Ch aims to keep AI interactions explicit, transparent, and user-driven, empowering developers to use LLM capabilities without leaving the terminal or surrendering control.
Highlights:
- Single Go binary: Starts instantly, streams responses in real time.
- Multi-provider support: OpenAI, Amazon Bedrock, Anthropic, Gemini, Groq, xAI, Mistral, Together, DeepSeek, and Ollama.
- File + web input: Load PDFs, DOCX, images (OCR), scrape URLs, or analyze YouTube content.
- Unix-style integration: Pipe command outputs or files directly; colors auto-suppressed for clean pipelines.
- Developer-focused: Record shell sessions as context, export code snippets, integrate with editors or other AI CLIs.
- Transparent & explicit: No background actions or silent model switches.
Install in one line:
curl -fsSL https://raw.githubusercontent.com/MehmetMHY/ch/main/install.... | bash
Links:
- GitHub: https://github.com/MehmetMHY/ch
- Demo Video: https://www.youtube.com/watch?v=AH0xG1iStf4
- Website: https://mehmetmhy.github.io/ch
I'd love feedback on the UX design, supported platforms, and ideas for new enhancements and features. Thank you for reading and I'm happy to answer any questions!
Best regards,
Mehmet Yilmaz
