Continue
Continue is an open-source AI programming assistant plugin for mainstream code editors including VS Code and JetBrains IDEs, designed to integrate large language model capabilities into development workflows. It supports code auto-completion, conversational code generation and editing, and codebase question answering. Users can configure connections to multiple AI models and external tool services, enabling developers to leverage AI capabilities directly within their editor to improve programming efficiency.
Core Facts
For applications with high call volumes, the total cost of self-hosted open-source model deployment is often significantly lower than using commercial APIs.
Mainstream development tools such as Cursor, Windsurf, and VS Code already support the MCP protocol.
Continue is an open-source VS Code/JetBrains plugin that supports connecting to multiple model backends including OpenAI, Anthropic, and local Ollama.
Timeline (last 90 days)
MCP allows projects like REA, as an npm package, to be plug-and-play for MCP-supporting clients such as Claude Desktop, Cursor and Continue.
Continue, Cline and Aider are plugins that support multiple backends and can switch models by task type within the same editor.
Regular Claude.ai web users currently cannot directly use the Instance Level MCP connection method and need to use Claude Code or other tools that support the MCP client protocol (such as Cursor or Continue)
Continue warns that its default context settings may exceed what a typical Ollama environment can handle.
The local AI programming pipeline consists of three types of roles: the editor that hosts the assistant, the runner that loads the model, and the agent that decides on actions.
The combination recommended by the blogger is Ollama + Continue + Qwen2.5-Coder (autocomplete) + Qwen3-Coder (agent tasks).
Continue's current configuration guide recommends using Qwen2.5-Coder 1.5B for the autocomplete role.
Open-source frontends and development tools such as Open WebUI, Continue, and Dify use Ollama as the preferred local inference runtime.
The main competitors in the open-source AI programming agent space include Aider, Continue, and OpenHands.
Aider focuses on command-line interactive programming, Continue takes the form of a VS Code/JetBrains plugin, and OpenHands, formerly known as OpenDevin, focuses on complex software engineering tasks.
22 more timeline events
All Facts (20)
For applications with high call volumes, the total cost of self-hosted open-source model deployment is often significantly lower than using commercial APIs.
80%VerifiedMainstream development tools such as Cursor, Windsurf, and VS Code already support the MCP protocol.
80%VerifiedContinue is an open-source VS Code/JetBrains plugin that supports connecting to multiple model backends including OpenAI, Anthropic, and local Ollama.
80%VerifiedUsing a routing mechanism to distribute requests to models of different scales based on task complexity can reduce the average call cost at the overall level.
75%VerifiedMainstream AI development tools including Claude, Cursor, and Zed all already support MCP.
75%VerifiedThe current AI programming tools market includes the IDE plugin layer (GitHub Copilot, Tabnine), the standalone editor layer (Cursor, Windsurf), the CLI Agent layer (Aider, Claude Code, Gemini CLI), and the framework layer (Cline, Continue).
70%VerifiedThe API aggregation layer is typically built on the OpenAI-compatible interface specification, allowing existing tools to connect simply by modifying the endpoint address.
65%VerifiedApplications that rely on the OpenAI SDK can switch to running locally on Ollama by modifying just one line of base_url configuration.
65%VerifiedIDE plugins such as Cursor, Continue, and Cline natively support the OpenAI-compatible interface format.
65%UnverifiedDeepSeek可以集成到VS Code、JetBrains系列IDE中，通过Continue、Codeium等插件实现智能代码补全、代码审查、文档自动生成等功能
85%UnverifiedContinue是一个开源AI编程助手框架，允许开发者自由选择后端模型，包括本地部署的Code Llama、DeepSeek Coder等开源模型
85%Unverified终端AI工具赛道在2024-2025年间经历了爆发式增长，参与者包括GitHub Copilot CLI、Gemini CLI、Anthropic的Claude Code、Cursor的终端集成以及Aider、Continue等开源项目
75%UnverifiedMCP allows projects like REA, as an npm package, to be plug-and-play for MCP-supporting clients such as Claude Desktop, Cursor and Continue.
50%UnverifiedContinue, Cline and Aider are plugins that support multiple backends and can switch models by task type within the same editor.
50%UnverifiedRegular Claude.ai web users currently cannot directly use the Instance Level MCP connection method and need to use Claude Code or other tools that support the MCP client protocol (such as Cursor or Continue)
50%UnverifiedContinue's current configuration guide recommends using Qwen2.5-Coder 1.5B for the autocomplete role.
50%UnverifiedThe combination recommended by the blogger is Ollama + Continue + Qwen2.5-Coder (autocomplete) + Qwen3-Coder (agent tasks).
50%UnverifiedThe local AI programming pipeline consists of three types of roles: the editor that hosts the assistant, the runner that loads the model, and the agent that decides on actions.
50%UnverifiedContinue warns that its default context settings may exceed what a typical Ollama environment can handle.
50%UnverifiedOpen-source frontends and development tools such as Open WebUI, Continue, and Dify use Ollama as the preferred local inference runtime.
50%