Is your feature request related to a problem? Please describe.
Currently, the open-webui project does not include support for the GenericOpenAi LLM provider. It would be beneficial for users to have the option to use this provider, especially those who are already familiar with it in other projects such as Mintplex-Labs/anything-llm.
Describe the solution you'd like
I would like to request the inclusion of the GenericOpenAi LLM provider in the open-webui project. This would include setting environment variables to define the base path (GENERIC_OPEN_AI_BASE_PATH), the API key (GENERIC_OPEN_AI_API_KEY), model preferences (GENERIC_OPEN_AI_MODEL_PREF), and token limits (GENERIC_OPEN_AI_MODEL_TOKEN_LIMIT), as implemented in the Mintplex-Labs/anything-llm project.
Describe alternatives you've considered
As an alternative, other LLM providers offering similar functionalities could be considered. However, GenericOpenAi is widely used and its inclusion would enhance the flexibility and usability of open-webui.
Additional context
Implementation references in the Mintplex-Labs/anything-llm project:
Originally created by @actus7 on GitHub (Oct 1, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/5846
**Is your feature request related to a problem? Please describe.**
Currently, the `open-webui` project does not include support for the GenericOpenAi LLM provider. It would be beneficial for users to have the option to use this provider, especially those who are already familiar with it in other projects such as `Mintplex-Labs/anything-llm`.
**Describe the solution you'd like**
I would like to request the inclusion of the GenericOpenAi LLM provider in the `open-webui` project. This would include setting environment variables to define the base path (`GENERIC_OPEN_AI_BASE_PATH`), the API key (`GENERIC_OPEN_AI_API_KEY`), model preferences (`GENERIC_OPEN_AI_MODEL_PREF`), and token limits (`GENERIC_OPEN_AI_MODEL_TOKEN_LIMIT`), as implemented in the `Mintplex-Labs/anything-llm` project.
**Describe alternatives you've considered**
As an alternative, other LLM providers offering similar functionalities could be considered. However, GenericOpenAi is widely used and its inclusion would enhance the flexibility and usability of `open-webui`.
**Additional context**
Implementation references in the `Mintplex-Labs/anything-llm` project:
- [server/utils/helpers/updateENV.js](https://github.com/Mintplex-Labs/anything-llm/blob/348d9c828565a56ae9b9e496171e06dd98f5981f/server/utils/helpers/updateENV.js#L164-L270)
- [server/utils/AiProviders/genericOpenAi/index.js](https://github.com/Mintplex-Labs/anything-llm/blob/348d9c828565a56ae9b9e496171e06dd98f5981f/server/utils/AiProviders/genericOpenAi/index.js#L1-L92)
- [frontend/src/components/LLMSelection/GenericOpenAiOptions/index.jsx](https://github.com/Mintplex-Labs/anything-llm/blob/348d9c828565a56ae9b9e496171e06dd98f5981f/frontend/src/components/LLMSelection/GenericOpenAiOptions/index.jsx#L1-L85)
- [server/utils/EmbeddingEngines/genericOpenAi/index.js](https://github.com/Mintplex-Labs/anything-llm/blob/348d9c828565a56ae9b9e496171e06dd98f5981f/server/utils/EmbeddingEngines/genericOpenAi/index.js#L1-L95)
I don't see any explanation for why we need another OpenAI middleware when we already have a functioning one that works with generic OpenAI API specification services? Am I missing something?
<!-- gh-comment-id:2385742942 -->
@justinh-rahb commented on GitHub (Oct 1, 2024):
I don't see any explanation for _why_ we need another OpenAI middleware when we already have a functioning one that works with generic OpenAI API specification services? Am I missing something?
@justinh-rahb I would like to explain the new feature we are developing, which focuses on integrating models available on the GitHub Marketplace, similar to the approach used by AnythingLLM. This feature aims to optimize the interaction and use of different machine learning models, providing a more cohesive and efficient experience for users. Read the Issue again carefully and you will understand.
<!-- gh-comment-id:2385996497 -->
@actus7 commented on GitHub (Oct 1, 2024):
@justinh-rahb I would like to explain the new feature we are developing, which focuses on integrating models available on the GitHub Marketplace, similar to the approach used by AnythingLLM. This feature aims to optimize the interaction and use of different machine learning models, providing a more cohesive and efficient experience for users. Read the Issue again carefully and you will understand.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @actus7 on GitHub (Oct 1, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/5846
Is your feature request related to a problem? Please describe.
Currently, the
open-webuiproject does not include support for the GenericOpenAi LLM provider. It would be beneficial for users to have the option to use this provider, especially those who are already familiar with it in other projects such asMintplex-Labs/anything-llm.Describe the solution you'd like
I would like to request the inclusion of the GenericOpenAi LLM provider in the
open-webuiproject. This would include setting environment variables to define the base path (GENERIC_OPEN_AI_BASE_PATH), the API key (GENERIC_OPEN_AI_API_KEY), model preferences (GENERIC_OPEN_AI_MODEL_PREF), and token limits (GENERIC_OPEN_AI_MODEL_TOKEN_LIMIT), as implemented in theMintplex-Labs/anything-llmproject.Describe alternatives you've considered
As an alternative, other LLM providers offering similar functionalities could be considered. However, GenericOpenAi is widely used and its inclusion would enhance the flexibility and usability of
open-webui.Additional context
Implementation references in the
Mintplex-Labs/anything-llmproject:@justinh-rahb commented on GitHub (Oct 1, 2024):
I don't see any explanation for why we need another OpenAI middleware when we already have a functioning one that works with generic OpenAI API specification services? Am I missing something?
@actus7 commented on GitHub (Oct 1, 2024):
@justinh-rahb I would like to explain the new feature we are developing, which focuses on integrating models available on the GitHub Marketplace, similar to the approach used by AnythingLLM. This feature aims to optimize the interaction and use of different machine learning models, providing a more cohesive and efficient experience for users. Read the Issue again carefully and you will understand.
@justinh-rahb commented on GitHub (Oct 1, 2024):
I've read it thrice, I remain unconvinced. Please, elaborate.