Suggestions & Feature Requests for Archon

[NeoTech2079]

Hey Archon Devs,

I’ve been testing Archon pretty heavily and wanted to share some suggestions/requests. Not sure if some of these are already on the roadmap, but here’s what would really improve the workflow:

🧠 Model & Agent Management

Agent model dropdowns – Right now, for Gemini/ChatGPT (and soon Ollama), you have to know and manually input model names. It would be much smoother if selecting an Agent auto-populated a dropdown of supported Chat + Embedding models, so users can just pick without guesswork.

🌐 Web Crawling Controls

Whitelist/Blacklist support – Would love finer control over crawling. Ability to exclude entire directories (/bio, /contact, /careers) or entire domains (e.g. github.com). This would cut down noise and focus crawls on valuable content.

📚 Knowledge Base Improvements

Better grouping UI/UX – Right now you can group items, but:

Once grouped, you can’t easily add more items without recreating.

You can only remove items one by one (tedious).

Group names aren’t clearly displayed or clickable.
Suggestion: let us see the group name, click to expand a list of items, add/remove multiple at once.

⚡ Embedding Enhancements

Parallel embeddings – If feasible, allow embeddings to be split across multiple backends (e.g. Gemini + ChatGPT simultaneously) to reduce processing time.

Code detection in docs – Uploads (PDF, MD, DOC) often fail to detect code snippets properly, even when Markdown shows fenced code blocks. Improving this parsing would unlock way better code search and retrieval.

Closing note:
Archon is already shaping up to be powerful, and these additions would really push it into “next-level dev tool” territory.

Thanks for considering 🙏

[coleam00]

Thanks for the kind words and the suggestions for Archon too! We've got you covered with the model management, work on that already under way! Same with the web crawling controls, though that will be more future (next few months).

Feel free to make a feature request in the issues for the knowledge base improvements!

For embeddings - parallel embeddings is rather complex so I wouldn't want to do this now, but it is a good idea for the future.