3 ms·

Are there any good LLM plugins for Obsidian, beyond just throwing Claude Code / Codex at your markdown folder?

eightysixfour 1y ago
I put Claude Code in MCP server mode, then use it as an MCP server for claude desktop to interact with that folder. Works better for me since Claude Code desperately wants to edit things instead of chat sometimes.
gavmor 1y ago
There's no decent RAG functionality, AFAICT, but the Text-Generator plugin has been fantastic w/ larger contexts and a template that pulls either/both links and backlinks into the inference query window. Hands-down my most productive interface to LLMs for [years since GPT3.5] years running.
- Noumenon72 1y ago
  The fact there is no app to do "RAG on localhost for my own notes" really makes me wonder whether the tech works at all.
- obsidianbases1 1y ago
  > There's no decent RAG functionality Do you have any examples of what decent RAG functionality might look like? And where the current plugins fall short?
  - gavmor 1y ago
    I simply haven't gotten Smart Connections to index my vaults successfully, for whatever reason. So I can't give them a proper review. I'd hope that any inference performed in a given note would take as context not only the full text of the note, but its nearest neighbors—although, frankly, what I really want are the implicit definitions of proper nouns, ie links. I also frequently generate articles based on the subject's backlinks, ie mentions in other notes. Both of those functionalities are provided by Obsidian's core featureset, so all that's missing—and what might provide outsized value—is indexing notes based not only on the embeddings of their literal text, but on meta-text such as: the questions they might answer; the definitions of their principle subjects; counterfactual statements they falsify. All of which would allow the editor to bring up a relevant note when writing/reviewing another–ie the "unknown knowns". I think this technique is akin to "knowledge enrichment", or "meta-embeddings," but I think actually those terms have lower-level meanings. Of course, generating these embeddings would make the indexing process slower and more failure prone, and so I'd be even more likely to fail to get it working and call it a dud.
    - obsidianbases1 1y ago
      > is indexing notes based not only on the embeddings of their literal text, but on meta-text such as: the questions they might answer; the definitions of their principle subjects; counterfactual statements they falsify. That would be perfect! Smart Connections does seem to have improved a lot recently.
obsidianbases1 1y ago
Copilot, Smart Connections and Text Generator are all popular LLM plugins in Obsidian https://youtu.be/mZ8TJ59Hj28 https://youtu.be/mZ8TJ59Hj28 https://youtu.be/7Rvl9Sl29Jk https://youtu.be/7Rvl9Sl29Jk

I put Claude Code in MCP server mode, then use it as an MCP server for claude desktop to interact with that folder. Works better for me since Claude Code desperately wants to edit things instead of chat sometimes.

gavmor 1y ago

There's no decent RAG functionality, AFAICT, but the Text-Generator plugin has been fantastic w/ larger contexts and a template that pulls either/both links and backlinks into the inference query window. Hands-down my most productive interface to LLMs for [years since GPT3.5] years running.

Noumenon72 1y ago
The fact there is no app to do "RAG on localhost for my own notes" really makes me wonder whether the tech works at all.
obsidianbases1 1y ago
> There's no decent RAG functionality Do you have any examples of what decent RAG functionality might look like? And where the current plugins fall short?
- gavmor 1y ago
  I simply haven't gotten Smart Connections to index my vaults successfully, for whatever reason. So I can't give them a proper review. I'd hope that any inference performed in a given note would take as context not only the full text of the note, but its nearest neighbors—although, frankly, what I really want are the implicit definitions of proper nouns, ie links. I also frequently generate articles based on the subject's backlinks, ie mentions in other notes. Both of those functionalities are provided by Obsidian's core featureset, so all that's missing—and what might provide outsized value—is indexing notes based not only on the embeddings of their literal text, but on meta-text such as: the questions they might answer; the definitions of their principle subjects; counterfactual statements they falsify. All of which would allow the editor to bring up a relevant note when writing/reviewing another–ie the "unknown knowns". I think this technique is akin to "knowledge enrichment", or "meta-embeddings," but I think actually those terms have lower-level meanings. Of course, generating these embeddings would make the indexing process slower and more failure prone, and so I'd be even more likely to fail to get it working and call it a dud.
  - obsidianbases1 1y ago
    > is indexing notes based not only on the embeddings of their literal text, but on meta-text such as: the questions they might answer; the definitions of their principle subjects; counterfactual statements they falsify. That would be perfect! Smart Connections does seem to have improved a lot recently.

obsidianbases1 1y ago

Copilot, Smart Connections and Text Generator are all popular LLM plugins in Obsidian https://youtu.be/mZ8TJ59Hj28 https://youtu.be/mZ8TJ59Hj28 https://youtu.be/7Rvl9Sl29Jk https://youtu.be/7Rvl9Sl29Jk