7 Pull requests merged by 7 people

• Fix PropmptTemplate typo in structured-output-api.jpg

  #3304 merged Jul 17, 2025

• fix method tool call throws same method name

  #2774 merged Jul 17, 2025

• Extract VectorStoreRetriever interface from VectorStore

  #3827 merged Jul 17, 2025

• docs(mcp): add dynamic MCP client selection example

  #3751 merged Jul 16, 2025

• docs: update mcp specification link

  #3810 merged Jul 16, 2025

• Fix exmaple code for "Streaming Responses"

  #3802 merged Jul 16, 2025

• Fix McpServerAutoConfigurationIT

  #3792 merged Jul 11, 2025

15 Pull requests opened by 12 people

• Feat(rerank): Add Cohere Reranker with topN filtering and fallback support

  #3791 opened Jul 11, 2025

• McpSyncServer: set immediate execution when in a servlet context

  #3794 opened Jul 11, 2025

• feat(zhipuai): Support glm-4.1v-thinking-flash model

  #3796 opened Jul 11, 2025

• Feature (template): Supports multi character template separators

  #3801 opened Jul 12, 2025

• GH-2947: Add @ToolParam annotation support for parameter name binding

  #3803 opened Jul 13, 2025

• GH-3786: Provide `encodingFormat=base64` support for the OpenAI embedding model

  #3804 opened Jul 13, 2025

• GH-3805: Add title attribute support to MCP @Tool annotation

  #3806 opened Jul 13, 2025

• GH-3807: Use builder pattern instead of deprecated constructors in class McpToolUtils

  #3808 opened Jul 13, 2025

• fix: GH-3690, enhance RedisVectorStore with RedisVectorStoreBuilderCustomizer

  #3809 opened Jul 13, 2025

• GH-3657: Fix DeepSeek tool call content null issue

  #3817 opened Jul 14, 2025

• Fixes #3693 - Support latest Google Unifed SDK to access Google AI Models

  #3820 opened Jul 14, 2025

• feat: MCP tools support ToolContext

  #3831 opened Jul 16, 2025

• Fix GraalVM native compilation issue with Java 22

  #3833 opened Jul 17, 2025

• Add More AWS Timeout Options

  #3836 opened Jul 17, 2025

• Use maven wrapper and batch-mode in GitHub Actions

  #3841 opened Jul 17, 2025

16 Issues closed by 4 people

• Misspelling in Documentation

  #3682 closed Jul 17, 2025

• Fix PropmptTemplate typo in structured-output-api.jpg

  #3839 closed Jul 17, 2025

• Diagram structured-output-api PropmptTemplate Typo

  #3295 closed Jul 17, 2025

• MCP Server @Tool throws "Multiple tools with the same name"

  #3837 closed Jul 17, 2025

• MCP Server @Tool throws "Multiple tools with the same name"

  #2771 closed Jul 17, 2025

• Add MCP Client Streamable HTTP transport auto-configuration

  #3782 closed Jul 17, 2025

• VectorStore interface should be split, with readonly operations becoming a super interface

  #1290 closed Jul 17, 2025

• docs(mcp): add dynamic MCP client selection example

  #3830 closed Jul 16, 2025

• docs: update mcp specification link

  #3829 closed Jul 16, 2025

• Fix exmaple code for "Streaming Responses"

  #3828 closed Jul 16, 2025

• Internationalization and Localization of resources

  #1578 closed Jul 16, 2025

• QuestionAnswerAdvisor should be able to inject the document id and metadata into the prompt.

  #1294 closed Jul 16, 2025

• Make the setting spring.ai.openai.api-key optional

  #3753 closed Jul 15, 2025

• Doc: Wrong dependency info in the RAG doc

  #3798 closed Jul 15, 2025

• I hope the next version can provide the functionality of openai request and response formats

  #3560 closed Jul 14, 2025

• could not found io.modelcontextprotocol.sdk:mcp:jar:0.11.0-SNAPSHOT in repo

  #3789 closed Jul 11, 2025

24 Issues opened by 21 people

• Oracle Vector Store - METADATA IGNORED

  #3838 opened Jul 17, 2025

• AWS Bedrock, Increased Latency with Large Requests Exceeds Default Timeouts

  #3835 opened Jul 17, 2025

• Vector store support for S3-Vectors

  #3834 opened Jul 17, 2025

• Tool annotation name validation

  #3832 opened Jul 16, 2025

• ES 混合检索 IN语句失败 search_phase_execution_exception

  #3826 opened Jul 16, 2025

• Neo4j ChatMemoryRepository throw Exception: Invalid input '$'

  #3825 opened Jul 16, 2025

• [EPIC] - Support latest Google Unifed SDK to access Google AI Embedding Models

  #3824 opened Jul 15, 2025

• Feature Request: Tool Calling Pattern Analysis and Optimization

  #3822 opened Jul 15, 2025

• 在Ubuntu 22.04.4 LTS使用Spring AI调用GUPtack中qwen3-embedding-8b模型报HTTP 400 - Invalid HTTP request received.异常

  #3821 opened Jul 15, 2025

• MCP Server: Circular Reference when setting ToolCallbackProvider when ChatClient is used for Tools

  #3819 opened Jul 14, 2025

• Document ID is not supported when adding embeddings to AWS OpenSearch

  #3818 opened Jul 14, 2025

• ChatModel streaming chat response data is too large.

  #3816 opened Jul 14, 2025

• Startup error when mcp server and mcp client are in the same spring boot project

  #3815 opened Jul 14, 2025

• deepseek ResponseFormat missing default constructor

  #3814 opened Jul 14, 2025

• Enhance Spring AI Tool Calling with @RequiresConsent

  #3813 opened Jul 14, 2025

• The prefixedToolName method of the McpToolUtils class limits its length to 64 bits

  #3812 opened Jul 14, 2025

• VertexAI when transport=REST It sometimes happens Read timed out

  #3811 opened Jul 14, 2025

• Use builder pattern instead of deprecated constructors in class McpToolUtils

  #3807 opened Jul 13, 2025

• Add title attribute support to MCP @Tool annotation

  #3805 opened Jul 13, 2025

• StTemplateRenderer delimiter token allows multiple characters

  #3800 opened Jul 12, 2025

• Feature Request: Allow configuration of payload content key for QdrantVectorStore

  #3797 opened Jul 11, 2025

• open ai + vllm 工具调用模型死循环不返回

  #3793 opened Jul 11, 2025

• Add ChatClient getChatModel() method

  #3790 opened Jul 11, 2025

• response.getEvalDuration(); threw a NullPointerException.

  #3788 opened Jul 11, 2025

64 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Fix string return conversion in DefaultToolCallResultConverter

  #3756 commented on Jul 17, 2025 • 3 new comments

• Create new ToolChatOptions if input request has tool configuration without options

  #3397 commented on Jul 17, 2025 • 2 new comments

• Fix MCP protocol, related commands cannot be executed under Windows, such as npx

  #3323 commented on Jul 17, 2025 • 1 new comment

• ToolCallingChatOptions support internalToolExecutionMaxIterations, to limit the maximum number of tool calls and prevent infinite recursive calls to LLM in special cases

  #3380 commented on Jul 17, 2025 • 1 new comment

• add aws bedrock prompt-caching support

  #3213 commented on Jul 11, 2025 • 0 new comments

• feat(memory): First commit of file-based memory repository

  #3210 commented on Jul 11, 2025 • 0 new comments

• feat: support Qwen models provided by Alibaba Cloud

  #2835 commented on Jul 17, 2025 • 0 new comments

• Create MongoDB chat memory implementation

  #2679 commented on Jul 11, 2025 • 0 new comments

• feat: Add Mistral AI OCR API integration

  #2404 commented on Jul 11, 2025 • 0 new comments

• Add MultiVectorRetriever as a foundation for ParentDocumentRetriever

  #2349 commented on Jul 16, 2025 • 0 new comments

• feat(redis): Add Redis-based semantic caching and chat memory implementations

  #2295 commented on Jul 17, 2025 • 0 new comments

• Support StabilityAi Image Edit Api

  #1344 commented on Jul 11, 2025 • 0 new comments

• Added new VertexAI Anthropic's model module

  #1213 commented on Jul 11, 2025 • 0 new comments

• feat: add hugging face text to image integration

  #1162 commented on Jul 11, 2025 • 0 new comments

• Add self-query advisor

  #1019 commented on Jul 11, 2025 • 0 new comments

• Provide the ability to configure OpenAI client read timeout

  #365 commented on Jul 16, 2025 • 0 new comments

• Model incompatible request argument supplied: temperature

  #3253 commented on Jul 17, 2025 • 0 new comments

• spring ai supports tool chaining

  #3691 commented on Jul 17, 2025 • 0 new comments

• feat(rag:etl): Add custom template support to KeywordMetadataEnricher

  #3252 commented on Jul 17, 2025 • 0 new comments

• Provides a Jinja2 implementation for TemplateRenderer

  #3259 commented on Jul 11, 2025 • 0 new comments

• Add Spring SpEL implementation of TemplateRenderer

  #3262 commented on Jul 11, 2025 • 0 new comments

• Refactor builder pattern in VertexAiGeminiSafetySetting

  #3326 commented on Jul 17, 2025 • 0 new comments

• fix(ai-rag): Fix query expansion result handling logic, same as #3345

  #3347 commented on Jul 17, 2025 • 0 new comments

• Added support for the "think" for Ollama

  #3386 commented on Jul 18, 2025 • 0 new comments

• fix: Incorrect order when Advisors have the same order

  #3461 commented on Jul 17, 2025 • 0 new comments

• GH-3540: Allow user-provided embeddings in VectorStore

  #3541 commented on Jul 15, 2025 • 0 new comments

• enhanced-deepseek cot

  #3703 commented on Jul 11, 2025 • 0 new comments

• feat: support the chat_template_kwargs, with OpenAI-Compatible Server…

  #3744 commented on Jul 11, 2025 • 0 new comments

• Support responseSchema in VertexAiGeminiChatOptions

  #3765 commented on Jul 14, 2025 • 0 new comments

• refactor: Simplifies VectorStoreChatMemoryAdvisor code

  #3773 commented on Jul 17, 2025 • 0 new comments

• fix: Handling default implementation of ToolCallback#call(String,ToolContext)

  #3784 commented on Jul 13, 2025 • 0 new comments

• fix: Fixed the issue where the returnDirect attribute was not effective on the MCP server side.

  #3787 commented on Jul 11, 2025 • 0 new comments

• openai options support extra params

  #3409 commented on Jul 11, 2025 • 0 new comments

• com.google.api.gax.rpc.UnknownException: java.net.SocketTimeoutException: Read timed out

  #1577 commented on Jul 11, 2025 • 0 new comments

• Spring AI MCP Streamable HTTP

  #3145 commented on Jul 11, 2025 • 0 new comments

• Add Support for OpenAI based64 format embedding responses

  #3786 commented on Jul 11, 2025 • 0 new comments

• RedisVectorStore auto-configuration does not support metadataFields, causing chat memory retrieval failure

  #3690 commented on Jul 11, 2025 • 0 new comments

• Reasoning options documented in docs

  #3725 commented on Jul 11, 2025 • 0 new comments

• Can this Spring AI compete official Typescript AI SDKs?

  #3750 commented on Jul 11, 2025 • 0 new comments

• Support A2A Protocol

  #2911 commented on Jul 11, 2025 • 0 new comments

• I need a solution to instable tool calling behavior in deepseek

  #3602 commented on Jul 11, 2025 • 0 new comments

• Request to Make ChatOptions Runtime-Validated Instead of Required at Startup

  #3404 commented on Jul 11, 2025 • 0 new comments

• After the mcp server restarts, the client will return 404.

  #3337 commented on Jul 11, 2025 • 0 new comments

• Tools MethodToolCallback， parameter name cannot be matched

  #2947 commented on Jul 12, 2025 • 0 new comments

• Support accessing HTTP headers in ToolContext when using WebFlux

  #2757 commented on Jul 12, 2025 • 0 new comments

• ChatClient streaming does not work with LM Studio Local LLM API Server

  #2445 commented on Jul 13, 2025 • 0 new comments

• Unable to use spring ai with LMStudio using spring-ai openai module

  #2441 commented on Jul 13, 2025 • 0 new comments

• block()/blockFirst()/blockLast() are blocking, which is not supported in thread reactor-http-nio-3

  #2533 commented on Jul 14, 2025 • 0 new comments

• server.servlet.context-path not properly factoring into SSE MCP server path

  #2512 commented on Jul 14, 2025 • 0 new comments

• Cannot invoke "String.equals(Object)" because "toolName" is null

  #2813 commented on Jul 14, 2025 • 0 new comments

• JsonParser is using a self instantiated ObjectMapper, allow usage of a custom ObjectMapper instance

  #2494 commented on Jul 15, 2025 • 0 new comments

• ToolContext is not supported by the MCP tools

  #3505 commented on Jul 15, 2025 • 0 new comments

• Official Claude 4 Support

  #3783 commented on Jul 15, 2025 • 0 new comments

• Add LiteLLM support

  #2734 commented on Jul 15, 2025 • 0 new comments

• (Feature request) Add support for latency optimized performanceConfig (AWS bedrock converse)

  #3032 commented on Jul 15, 2025 • 0 new comments

• MCP server: Authentication lost in tool execution

  #2506 commented on Jul 16, 2025 • 0 new comments

• Support of Guardrails in Spring AI

  #1081 commented on Jul 16, 2025 • 0 new comments

• Optimized aggregation advisors for streaming scenarios

  #1439 commented on Jul 16, 2025 • 0 new comments

• Provider Exception Failure Analysis abstraction to provide better error messages to users

  #1704 commented on Jul 16, 2025 • 0 new comments

• Please confirm if it is a bug,Springai encountered an error while calling VLLM

  #3775 commented on Jul 16, 2025 • 0 new comments

• Streaming with tools not working with Anthropic ChatClient

  #3767 commented on Jul 16, 2025 • 0 new comments

• native compilation under java 22 fails for an example that works on java 21

  #494 commented on Jul 17, 2025 • 0 new comments

• Question about sglang + deepseek

  #3657 commented on Jul 17, 2025 • 0 new comments

• Spring AI fails on second vLLM call with 400 due to malformed tool_calls

  #3438 commented on Jul 17, 2025 • 0 new comments