Aurora Mobile Limited has announced a beta plugin connecting Modellix.ai with the fast-growing DeepSeek Harness AI coding agent.
The new plugin, dsh-modellix, enables users to use multiple LLMs via Modellix’s single gateway. Moreover, a chosen set of models is available for free use. Modellix.ai is the unified platform of Aurora Mobile for AI media generation. The beta version of the plugin allows integration with DeepSeek Harness’s model selection mechanism.
DeepSeek Harness was released as an open-source project on August 13, 2026. Since its release, the AI code assistant has quickly garnered popularity among developers due to its plugin-centric design. This platform operates on the philosophy that “everything is a plugin.” This enables developers to extend their coding processes via third-party service integration.
A Unified Gateway for AI Coding Models
The DeepSeek Harness plugin helps users use the “modellix-ai/free-llm” option to use free models. It automatically sends all the requests to the best available open source LLMs. Also, users can choose the “zai/glm-4.7-flash” model without any cost per call. For usage, all developers require an account on Modellix with an API key. In addition, users can try more than 28 models without changing the API keys.
“AI coding agents are becoming the default front door to AI development, and the ecosystem around DeepSeek Harness is growing remarkably fast,” said Mr. Weidong Luo, Chairman and Chief Executive Officer of Aurora Mobile. “dsh-modellix is an early beta, and our goal is simple: a developer using Harness should be able to try the latest models including those billed at $0 in under two minutes, with one key that also unlocks image, video, and audio generation on the same platform. We’re shipping it early to learn what the community actually wants, and iterating from there.”
The plugin has three different abilities that operate independently: LLM, Web, and Design. The LLM ability helps to synchronize the live models of Modellix with the native model selection tool of Harness. The Web ability allows the web search and web fetch native tools to pass through the Modellix endpoint.
LLM features are still working during the beta stage. But, Design features are still previewing as Modellix is making compatibility between the model schema. Furthermore, Modellix uses openai and Anthropic protocols. Developers have the ability to route their SDKs and clients with a base url override feature.
These methods are compatible with the use of Codex, Claude Code, and Cursor without any modification to code. Moreover, the DeepSeek Harness allows users to have multiple AI vendors using a single gateway.
Want more marketing excellence stories? Visit MarTech News for insights, trends, and expert updates.
News Source: GlobeNewswire.com
