feat(models): add Atomic Chat local LLM provider plugin

[yanalialiuk]

Summary

• Add models/atomic_chat — a Dify model provider plugin for Atomic Chat
• Connects to Atomic Chat's OpenAI-compatible Local API Server (default http://127.0.0.1:1337/v1)
• Supports customizable LLM models with tool calling, streaming, and optional vision/thinking settings
• Includes setup docs for local installs and Docker (host.docker.internal)

Why a plugin (not dify core)

Dify model providers live in dify-official-plugins and are loaded by plugin-daemon. This follows the same pattern as ollama, lemonade, and openai_api_compatible.

Test plan

• Build/install plugin package in a Dify instance with plugin-daemon enabled
• Add a model with endpoint http://127.0.0.1:1337/v1 (or http://host.docker.internal:1337/v1 from Docker)
• Run a chat/workflow app using the configured Atomic Chat model
• Verify credential validation succeeds when Atomic Chat Local API Server is running

[gemini-code-assist]

Code Review

This pull request introduces the Atomic Chat model provider for Dify, enabling local LLM integration via an OpenAI-compatible API. The implementation includes custom credential validation, multimodal support, and thinking/reasoning stream filtering. However, several issues need to be addressed: potential KeyError and AttributeError exceptions when accessing optional credential fields like mode and endpoint_url, a potential initialization failure when api_key is absent, and a bug in the stream filtering logic that can leak blocks if they are split across chunks. Additionally, the _CommonOpenAI helper class is currently unused dead code and should be integrated or removed.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.