AutoDev Takes the Lead: Bi-directional Enabling of MCP Protocol, Building a New Ecosystem of AI Coding Tools

In the field of Artificial Intelligence coding, how to make AI Intelligent Bodies (Agents) more effectively utilize tools to complete complex software development tasks has been a core issue of great concern. "Tool Use/Function calling is a key technology born in this context. A perfect software development process, in addition to the integrated development environment (IDE) itself and its rich plug-in ecosystem, but also inseparable from a series of external tools such as version control system Git, containerization technology Docker, container orchestration platform Kubernetes, continuous integration/continuous delivery tools Jenkins and other collaborative operation. How to enable AI to perceive and skillfully utilize these tools has become the key to improving AI coding capabilities.

Recently, the AI coding tool AutoDev announced a major update centered around the implementation of bi-directional integration with the Model Context Protocol (MCP). This means that AutoDev can be used not only as a MCP It can also be used as an MCP client to call all kinds of tools from the MCP ecosystem. This innovative initiative undoubtedly provides new ideas for the future development direction of AI coding tools.

From Progressive AI Agent Programs to AutoDev as an MCP Service

In the larger scenario of automated computer use, IDEs are just one of many tools that can be called upon by AI Agents. Currently, the development of AI Agents is in an incremental stage. Traditional user interfaces (UIs) operate inefficiently, which limits the automation capabilities of AI Agents. Therefore, the IDE-centric AI coding model is still the mainstream, but the future trend is to move to an Agent Tool-centric model, where the AI Agent can manipulate the browser to obtain requirement information, launch the IDE to write code, call DevOps tools to publish the application, and even carry out a wider range of automation operations.

In the face of the wave of AI Agent tool calls, AutoDev has taken the lead in positioning itself as an MCP service. This means that any agent tool that supports the MCP protocol, such as Cursor, Cline, GitHub, etc., will be able to call the MCP service. Copilot etc., can directly invoke the services provided by AutoDev to access high-quality contextual information in the IDE environment. This shift breaks down barriers between tools and lays the foundation for a stronger AI coding ecosystem.

Eco to MCP service from MCP i.e. Agent Tool i.e. AutoDev command.

In the past, AutoDev focused on building an end-to-end AI-assisted development process by leveraging the IDE's ecosystem and plug-in system. As AutoDev integrates more and more plug-ins, the need for an Agent Tool has become more and more prominent.

Agent Tool Laying the Baseline of AI IDE Capabilities

A comparison of the many AI coding tools on the market reveals that VSCode-based tools are highly homogenized in terms of functionality, and the toolset they provide is often similar. However, AI coding plug-ins that are integrated into the JetBrains IDE platform, such as AutoDev and JetBrains Junie, show much greater capabilities, providing deeper IDE functionality including Abstract Syntax Tree (AST) analysis, code debugging, and Fully Qualified Name (FQN) lookup.

While VSCode-based AI coding tools typically offer about ten tools, the JetBrains platform's AI coding plug-ins offer about twenty, but this is just the tip of the iceberg in terms of the IDE's basic capabilities. As AI coding technology extends into broader areas such as requirements analysis, deployment, and operations and maintenance, the richness and specialization of the tool ecosystem becomes even more demanding.

MCP The Rise of the Open Source Ecosystem

The Model Context Protocol (MCP) is provided by the Anthropic (the company behind the Claude model), aims to provide a standardized set of interfaces for large-scale language model (LLM) applications, enabling them to easily access external information, tools, and resources. While AutoDev already has strong customization capabilities, such as support for custom agents, it still has limitations in the face of a rapidly iterating technology environment and diverse user needs:

• Unpredictability of tools: Many businesses or organizations have customized tools in-house that are difficult to pre-integrate into common platforms such as AutoDev.
• Booming MCP Ecosystem: The MCP protocol and its related tools ecosystem are rapidly developing and gradually becoming the industry standard. Especially after the introduction of the MCP protocol by new programming tools such as Cursor and Cline, a large number of MCP services and open source implementations have emerged overseas, building a vibrant ecosystem.

AutoDev × MCP: Two-Way Empowerment to Build an Open AI Coding Platform

Based on its deep understanding of AI coding trends and the value of the MCP protocol, AutoDev has innovatively launched the AutoDev × MCP two-way empowerment program. Specifically:

• AutoDev as MCP Server: AutoDev can act as an MCP service provider, opening up its features and services to any Agent Tool.
• AutoDev as an MCP client: AutoDev can act as a consumer of MCP services, invoking the rich set of tools and services in the MCP ecosystem.

With these two models, AutoDev not only leverages the resources of the MCP ecosystem, but also lays a solid foundation for the expansion and openness of its capabilities.

AutoDev as an MCP Server: Open IDE Contexts

AutoDev is built with the ability to act as an MCP server based on the JetBrains platform's MCP solution (MCP functionality needs to be enabled manually). Users can configure it in JSON format as follows Cline Sample configuration in the plugin:

{
"mcpServers": {
"AutoDev": {
"command": "npx",
"args": [
"-y".
"@jetbrains/mcp-proxy"
],, "disabled": false, false, false, false, false, false, false, false, false
"disabled": false,
"autoApprove": []
}
}
}

In the current version, AutoDev is mainly based on the basic capabilities provided by the official, and initially extends some database-related tools. In the future, as the domestic MCP ecosystem matures and develops, AutoDev will further expand its capabilities as an MCP server.

AutoDev as an MCP Client: Extending the Tool Boundary

Users can configure the MCP service on the Custom Agent page of AutoDev. The following is an example of the file system service provided by MCP:

{
"mcpServers": {
"filesystem": {
"command": "npx",
"args": [
"-y".
"@modelcontextprotocol/server-filesystem",
"/Volumes/source/ai/auto-dev"
]
}
}
}

Once configured, the tools provided by this MCP service can be invoked in AutoDev. For example, a file system service may provide a list_directory tool that can be invoked in AutoDev in the form of a DevIns Command:

/list_directory
```json
{
"path":"/Volumes/source/ai/autocrud/docs/mcp"
}

With AutoDev's powerful DevIns Command feature, you have the flexibility to invoke any MCP service in AutoDev, and even intelligently orchestrate and utilize these tools through the Agent.

concluding remarks

AutoDev's two-way empowerment of the MCP protocol is not only a technical upgrade, but also an important leap in the development concept of AI coding tools. By embracing the MCP ecosystem, AutoDev breaks the closed nature of traditional AI coding tools and builds a more open, flexible and scalable platform. This indicates that future AI coding tools will focus more on ecological cooperation and tool interconnection, and jointly promote AI-assisted software development to new heights.