[GH-ISSUE #1269] [FEAT]: Open the OpenAI endpoint for customization for embedding modells #795

Closed
opened 2026-02-22 18:21:25 -05:00 by yindo · 4 comments
Owner

Originally created by @alexanderMal on GitHub (May 2, 2024).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/1269

What would you like to see?

Dear AnythingLLM team thank you for your quick responses and good work!
As the community is pretty new and still in its finding phase standards are rare and as of now only determined by OpenAI it seems. More and more endpoints support the OpenAI standard for example the huggingface text-embeddings-inference. As the AnythingLLM Application already provides the OpenAPI endpoint it might be very beneficial to allow a bit more customization like the URL for this endpoint. This would allow for a major growth of the compatible ecosystem and might be especially beneficial for embedding models.

My reason for requesting this feature is the rapid innovation in embedding models also focused on non-English speakers.

Keep up the good work!

P.S. As i mistakenly closed this feature without a comment:

If this also includes the embedding model than https://github.com/Mintplex-Labs/anything-llm/pull/1178 is the feature I would like to have. Yes!
I did not find this request before sorry for that and thank you for letting me know!

Originally created by @alexanderMal on GitHub (May 2, 2024). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/1269 ### What would you like to see? Dear AnythingLLM team thank you for your quick responses and good work! As the community is pretty new and still in its finding phase standards are rare and as of now only determined by OpenAI it seems. More and more endpoints support the OpenAI standard for example the [huggingface text-embeddings-inference](https://github.com/huggingface/text-embeddings-inference/tree/main). As the AnythingLLM Application already provides the OpenAPI endpoint it might be very beneficial to allow a bit more customization like the URL for this endpoint. This would allow for a major growth of the compatible ecosystem and might be especially beneficial for embedding models. My reason for requesting this feature is the rapid innovation in embedding models also focused on non-English speakers. Keep up the good work! P.S. As i mistakenly closed this feature without a comment: If this also includes the embedding model than https://github.com/Mintplex-Labs/anything-llm/pull/1178 is the feature I would like to have. Yes! I did not find this request before sorry for that and thank you for letting me know!
yindo added the enhancementstage: specificationsfeature request labels 2026-02-22 18:21:25 -05:00
yindo closed this issue 2026-02-22 18:21:25 -05:00
Author
Owner

@timothycarambat commented on GitHub (May 2, 2024):

Does the Generic OpenAi LLM provider not fill that gap at this time?
https://github.com/Mintplex-Labs/anything-llm/pull/1178

All the fields are open input and will just get used the OpenAI wrapper SDK - so this should be okay?

@timothycarambat commented on GitHub (May 2, 2024): Does the `Generic OpenAi` LLM provider not fill that gap at this time? https://github.com/Mintplex-Labs/anything-llm/pull/1178 All the fields are open input and will just get used the OpenAI wrapper SDK - so this should be okay?
Author
Owner

@alexanderMal commented on GitHub (May 6, 2024):

I just took a deeper look into feature #1178 and found this is not the feature I would like to have. My specific usecase is using the OpenAI kommunication protocol for other embedding models. Like a locally hosted huggingface text-embeddings-interface

@alexanderMal commented on GitHub (May 6, 2024): I just took a deeper look into feature #1178 and found this is not the feature I would like to have. My specific usecase is using the OpenAI kommunication protocol for other embedding models. Like a locally hosted huggingface text-embeddings-interface
Author
Owner

@Vladimir-Va commented on GitHub (May 17, 2024):

Hello gentlemens, I'm trying to configure Generic OpenAI and I get an error:
400 Invalid type for 'max_tokens': expected an integer, but got a string instead.
In the max_tokens field, except for numbers, nothing can be configured.

@Vladimir-Va commented on GitHub (May 17, 2024): Hello gentlemens, I'm trying to configure Generic OpenAI and I get an error: `400 Invalid type for 'max_tokens': expected an integer, but got a string instead.` In the max_tokens field, except for numbers, nothing can be configured.
Author
Owner

@timothycarambat commented on GitHub (May 18, 2024):

@Vladimir-Va resolved via https://github.com/Mintplex-Labs/anything-llm/pull/1445
(not in most recent desktop version until 1.5.5+)

@timothycarambat commented on GitHub (May 18, 2024): @Vladimir-Va resolved via https://github.com/Mintplex-Labs/anything-llm/pull/1445 (not in most recent desktop version until 1.5.5+)
yindo changed title from [FEAT]: Open the OpenAI endpoint for customization for embedding modells to [GH-ISSUE #1269] [FEAT]: Open the OpenAI endpoint for customization for embedding modells 2026-06-05 14:37:14 -04:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#795