Jump to content

HTML version of current docs? (for AI coder)

Recommended Posts

Is there perhaps still an HTML version around of the full and up-to-date Opusmodus documentation?



I am currently playing around with multiple AI coders (so far, I used Github Copilot, aider and Cursor at work). I  am still exploring, but they definitely are helpful, in particular for popular programming languages and libraries. Not so much for more niche environments like Opusmodus yet.


However, this is something that we can change, e.g., by using RAG (retrieval augmented generation, i.e. letting some system automatically extract user query-relevant information from project-relevant sources and then automatically complement the query with that additional information). I am reasonably confident from other experience that such an approach could make an AI coder with RAG support (e.g., Cursor) quite useful for Opusmodus development -- once we integrate relevant documentation in the loop. However, Cursor currently only supports HTML format for that purpose.


So, is there perhaps still an HTML version around of the full and up-to-date Opusmodus documentation? If not, no worries. I will find a way to automatically convert all the current PDF files to a suitable format.



PS: I should perhaps add that this approach not only allows to have some AI coder help with the actual coding, but you then can also quasi "talk" with the relevant documentation. Imagine some Opusmodus expert you can ask anything, and they immediately give some response based on some information somewhere buried in the documentation. Search only gets you so far, and this approach allows combining information from multiple parts of the documentation in a single response. I am "talking" with my own day-job code similarly  these days...  

Link to comment
Share on other sites

PPS: For cases where you want to talk with some docs and RAG is not good enough (i.e. where a semantic search across different docs might not be suitable, because, say, the system might need to first understand some terms defined in one place, before understanding the actual relevant parts of another text location) I found that an language model with a very large context window that can be fed your whole document can be an interesting alternative. For example, I have used Google NotebookLM (https://notebooklm.google/) to talk with a selection of music theory textbooks in the context window. 

Link to comment
Share on other sites

  • torstenanders changed the title to HTML version of current docs? (for AI coder)

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...

Important Information

Terms of Use Privacy Policy