🤖 feat: Support Google Agents, fix Various Provider Configurations#5126
Merged
Conversation
…upported by langchain, requires PR
owengo
pushed a commit
to openwengo/LibreChat
that referenced
this pull request
Apr 27, 2025
GOOGLE_EXTRA_HEADERS for GenAI 🛡️ feat: Google Reverse Proxy support, `CIVIC_INTEGRITY` harm category (danny-avila#5037) * 🛡️ feat: Google Reverse Proxy support, `CIVIC_INTEGRITY` harm category * 🔧 chore: Update @langchain/google-vertexai to version 0.1.4 in package.json and package-lock.json * fix: revert breaking Vertex AI changes --------- Co-authored-by: KiGamji <maloyh44@gmail.com> Enable google search / retrieval for google endpoints 🔐 feat: Implement Allowed Action Domains (danny-avila#4964) * chore: RequestExecutor typing * feat: allowed action domains * fix: rename TAgentsEndpoint to TAssistantEndpoint in typedefs * chore: update librechat-data-provider version to 0.7.62 🤖 feat: Support Google Agents, fix Various Provider Configurations (danny-avila#5126) * feat: Refactor ModelEndHandler to collect usage metadata only if it exists * feat: google tool end handling, custom anthropic class for better token ux * refactor: differentiate between client <> request options * feat: initial support for google agents * feat: only cache messages with non-empty text * feat: Cache non-empty messages in chatV2 controller * fix: anthropic llm client options llmConfig * refactor: streamline client options handling in LLM configuration * fix: VertexAI Agent Auth & Tool Handling * fix: additional fields for llmConfig, however customHeaders are not supported by langchain, requires PR * feat: set default location for vertexai LLM configuration * fix: outdated OpenAI Client options for getLLMConfig * chore: agent provider options typing * chore: add note about currently unsupported customHeaders in langchain GenAI client * fix: skip transaction creation when rawAmount is NaN 🔧 refactor: Improve Agent Context & Minor Fixes (danny-avila#5349) * refactor: Improve Context for Agents * 🔧 fix: Safeguard against undefined properties in OpenAIClient response handling * refactor: log error before re-throwing for original stack trace * refactor: remove toolResource state from useFileHandling, allow svg files * refactor: prevent verbose logs from axios errors when using actions * refactor: add silent method recordTokenUsage in AgentClient * refactor: streamline token count assignment in BaseClient * refactor: enhance safety settings handling for Gemini 2.0 model * fix: capabilities structure in MCPConnection * refactor: simplify civic integrity threshold handling in GoogleClient and llm * refactor: update token count retrieval method in BaseClient tests * ci: fix test for svg 📜 refactor: Log Error Messages when OAuth Fails (danny-avila#5337) 🔐 feat: Implement Allowed Action Domains (danny-avila#4964) * chore: RequestExecutor typing * feat: allowed action domains * fix: rename TAgentsEndpoint to TAssistantEndpoint in typedefs * chore: update librechat-data-provider version to 0.7.62 🛡️ feat: Google Reverse Proxy support, `CIVIC_INTEGRITY` harm category (danny-avila#5037) * 🛡️ feat: Google Reverse Proxy support, `CIVIC_INTEGRITY` harm category * 🔧 chore: Update @langchain/google-vertexai to version 0.1.4 in package.json and package-lock.json * fix: revert breaking Vertex AI changes --------- Co-authored-by: KiGamji <maloyh44@gmail.com> Enable google search / retrieval for google endpoints
MichielMAnalytics
pushed a commit
to MichielMAnalytics/ProAI
that referenced
this pull request
Jun 6, 2025
…anny-avila#5126) * feat: Refactor ModelEndHandler to collect usage metadata only if it exists * feat: google tool end handling, custom anthropic class for better token ux * refactor: differentiate between client <> request options * feat: initial support for google agents * feat: only cache messages with non-empty text * feat: Cache non-empty messages in chatV2 controller * fix: anthropic llm client options llmConfig * refactor: streamline client options handling in LLM configuration * fix: VertexAI Agent Auth & Tool Handling * fix: additional fields for llmConfig, however customHeaders are not supported by langchain, requires PR * feat: set default location for vertexai LLM configuration * fix: outdated OpenAI Client options for getLLMConfig * chore: agent provider options typing * chore: add note about currently unsupported customHeaders in langchain GenAI client * fix: skip transaction creation when rawAmount is NaN
mcbuckeye
pushed a commit
to mcbuckeye/LibreChat
that referenced
this pull request
Aug 27, 2025
…anny-avila#5126) * feat: Refactor ModelEndHandler to collect usage metadata only if it exists * feat: google tool end handling, custom anthropic class for better token ux * refactor: differentiate between client <> request options * feat: initial support for google agents * feat: only cache messages with non-empty text * feat: Cache non-empty messages in chatV2 controller * fix: anthropic llm client options llmConfig * refactor: streamline client options handling in LLM configuration * fix: VertexAI Agent Auth & Tool Handling * fix: additional fields for llmConfig, however customHeaders are not supported by langchain, requires PR * feat: set default location for vertexai LLM configuration * fix: outdated OpenAI Client options for getLLMConfig * chore: agent provider options typing * chore: add note about currently unsupported customHeaders in langchain GenAI client * fix: skip transaction creation when rawAmount is NaN
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
I added initial support for Google agents and improved the handling of LLM configurations in agent initialization. I also made various fixes and optimizations related to client options and message caching.
getLLMConfig.llmConfigfields, noting thatcustomHeadersare not currently supported by LangChain's GenAI Client.Other
chatV2controller to avoid caching empty responses.BaseClientandPluginsClient.Change Type
Checklist