TypeScript and Python SDK giving AI agents 1000+ pre-authenticated app toolkits, per-user sessions, and a sandbox to take actions.
pig-mesh/deepseek4j
A Java SDK for the DeepSeek API, with chat, function calling, structured JSON output, embeddings, and a Spring Boot starter.
- Stars
- 750
- Forks
- 142
- Open issues
- 16
- Last commit
- 13 days ago
What is it?
deepseek4j is a Java SDK for calling the DeepSeek API. It covers chat completions with reasoning output, function calling, JSON structured output, and embedding generation through an OpenAI-compatible protocol. The default model is DeepSeek Flash, and DeepSeek V4 Pro can be selected through configuration or per request.
For Spring Boot 2.x and 3.x projects, a starter module auto-configures the client from application.yml or application.properties, and streaming responses are exposed as Reactor Flux values for SSE endpoints. The core module can also be used without Spring by creating a client with DeepSeekClient.builder(). Solon is listed as a supported framework.
The SDK also wraps DeepSeek's official Search API for structured web results, and provides a helper that searches first and then streams a model answer grounded in those results.
Key features
- 01Chat completions with reasoning output and session billing data
- 02Function calling and JSON structured output
- 03Embedding generation via OpenAI-compatible API
- 04Spring Boot starter with auto-configured client and streaming support
- 05Official DeepSeek Search API wrapper with sync and async calls
- 06Configurable proxy, timeouts, and request/response logging
Use cases
- Adding DeepSeek chat features to a Spring Boot backend
- Streaming model answers to a browser over server-sent events
- Calling DeepSeek from a plain Java application without Spring
- Building search-grounded answers from official web search results
Quick start
- 1
Add the Maven dependency (Spring Boot starter, or deepseek4j-core without Spring Boot)
- 2
Set your DeepSeek API key in application.yml under deepseek.api-key
- 3
Inject DeepSeekClient and call chatFluxCompletion(prompt) to stream a reply
Things to know
- Requires a DeepSeek API key to call the service.
- Most detailed documentation is in Chinese.
- Version 1.6.0 contains breaking search API changes, so upgrading from 1.5.0 needs code updates.
Topics
FAQ
- What is deepseek4j?
- A Java SDK for the DeepSeek API, with chat, function calling, structured JSON output, embeddings, and a Spring Boot starter. deepseek4j is a Java SDK for calling the DeepSeek API. It covers chat completions with reasoning output, function calling, JSON structured output, and embedding generation through an OpenAI-compatible protocol. The default model is DeepSeek Flash, and DeepSeek V4 Pro can be selected through configuration or per request.
- Is deepseek4j free to use? What license does it have?
- deepseek4j is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is deepseek4j still maintained?
- Actively maintained. The last commit to deepseek4j was on Oct 25, 2026.
- What can I use deepseek4j for?
- Adding DeepSeek chat features to a Spring Boot backend Streaming model answers to a browser over server-sent events Calling DeepSeek from a plain Java application without Spring Building search-grounded answers from official web search results
- How do I install deepseek4j?
- Add the Maven dependency (Spring Boot starter, or deepseek4j-core without Spring Boot) Set your DeepSeek API key in application.yml under deepseek.api-key Inject DeepSeekClient and call chatFluxCompletion(prompt) to stream a reply
- Where can I get deepseek4j?
- The source code and installation instructions are on GitHub: https://github.com/pig-mesh/deepseek4j
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A collection of agent skills that give AI coding assistants up-to-date context for building apps with the Gemini API and SDKs.
A Java 8+ SDK for building AI agents, with unified access to multiple LLM platforms, tool calling, MCP, RAG, and a built-in coding agent CLI/TUI.
Experimental, educational Python framework for lightweight multi-agent orchestration built on Agents and handoffs over the Chat Completions API.
A Home Assistant custom component that lets an OpenAI-compatible LLM control your devices, call services, and create automations.
Python client library that wraps multiple AI provider APIs behind one simple interface for text, structured data, and conversations.