Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.qodo.ai/llms.txt

Use this file to discover all available pages before exploring further.

The Context Engine MCP is a standalone developer tool that exposes codebase intelligence via the Model Context Protocol. For the Context Engine as the core technology layer of the Qodo platform, see Understanding the Context Engine.
Qodo Context Engine MCP, formerly Qodo Aware, provides multi-repo intelligence via the Model Context Protocol. It is a deep research agent that is purpose-built for large codebases to surface organizational insights and help you solve complex technical tasks. Qodo Context Engine MCP brings deep, contextual intelligence to your workflow by letting you query your remote codebases, access organizational knowledge, and explore best practices, all without context switching. It combines agent-based orchestration, advanced retrieval techniques, and codebase indexing to function like a “Principal Engineer” embedded in your workflow. At its core, Qodo Context Engine MCP is a code understanding engine that powers the rest of the Qodo platform. It connects to your organization’s code across as many as thousands of repositories, and builds a rich internal map of your architecture, logic, and development practices. Then it uses AI agents to help you query that knowledge, solve problems, and plan confidently without switching tools or asking around. Think of it as an AI engineer that reads all your code, and can explain it back to you, reason through changes, or surface important relationships. Using the Qodo Context Engine MCP, you can:
  • Connect products like Claude Desktop, Cursor, or others directly to your remote codebase via Qodo Context Engine.
  • Get consistent, organization-aware responses grounded in your real repositories, services, and documentation.
  • Benefit from Qodo Context Engine’s agentic reasoning and advanced retrieval, even outside the Qodo interface.
  • Eliminate context switching. Your tools now understand your code as if a senior engineer were helping.

​
What can it do?

Qodo Context Engine MCP connects to your remote codebase, indexes it deeply, and answers your questions like a Principal Engineer would: fast when it can, thorough when it needs to be. Using Qodo Context Engine MCP, you can query your team’s remote codebases, access organizational knowledge, and follow guidelines, all without context switching. It can be integrated into any tool that supports the Model Context Protocol, such as Cursor, Windsurf, Claude Desktop, and GitHub Copilot. It supports multiple agents to fit your workflow. Use Ask Agent when you need quick understanding, and Deep Research Agent when you’re planning refactors, architectural changes, or debugging across systems. You can also scope answers to specific repositories.

​
Example questions you can ask

  • “Where is this function used across our repos?”
  • “If I change this API response format, what breaks?”
  • “How does session management work end-to-end?”
  • “What services call this shared utility?”
Qodo Context Engine MCP understands usage patterns, dependencies, and architectural intent. It’s more than just code search.

​
How does it work?

Qodo Context Engine MCP combines Retrieval-Augmented Generation (RAG) with agentic reasoning to deliver intelligent, context-rich responses to your queries.
  1. Indexing: Qodo Context Engine MCP indexes your repositories from supported Git providers, creating a structured, multi-layered understanding of your codebase.
  2. Retrieval: When you ask a question, Qodo Context Engine MCP retrieves relevant functions, docs, commit history, and architectural patterns.
  3. Agentic Reasoning: Specialized agents analyze the retrieved context to understand relationships, dependencies, and intent, much like a senior engineer would.
  4. Generation: The model then crafts a clear, accurate response grounded in that understanding.
This makes Qodo Context Engine MCP an intelligent assistant that reads, interprets, and reasons over your entire codebase to help with implementation, debugging, and architectural decisions. Traditional RAG systems retrieve a few snippets and reason only over those. If the right context isn’t pulled at the start, the answer can miss the mark. Qodo Context Engine MCP goes further. It deeply indexes your entire codebase (across thousands of repos and years of history) and applies agentic reasoning on top, so it can handle both quick “what does this function do?” lookups and complex “what breaks if we refactor this?” research without getting stuck on the wrong context.

​
Why is it different?

Most AI tools only look at a few files. Qodo Context Engine MCP understands your entire system, so it goes beyond autocomplete and one-off completions to help you understand how things work, refactor safely, and make changes with full context. It does this with:
  • Advanced retrieval pipelines that find the right context, not just nearby code
  • Multi-agent reasoning that breaks down complex tasks
  • Enterprise-scale indexing that supports thousands of repos and years of history
Qodo Context Engine MCP helps you with real problems like:
  • Understanding how a feature works across backend, frontend, and data layers
  • Planning architectural changes across teams
  • Finding internal examples and best practices
  • Getting accurate impact analysis before merging

​
Security and privacy

Qodo Context Engine MCP is built with enterprise-grade security and privacy in mind. It’s designed to provide intelligent, codebase-aware assistance, without compromising your data or control.

​
Data access and scope

  • Code access is explicit: Qodo Context Engine MCP only analyzes the repositories you explicitly connect and tag. No code is accessed without user or admin configuration.
  • Context is scoped: Tagged repositories define the scope of what Qodo Context Engine MCP can see and reason about. You remain in full control of what’s included.

​
Data handling

  • No training on customer code: Qodo Context Engine MCP does not use your code for training any foundation models.
  • Temporary retrieval only: Relevant code snippets are retrieved for context at query time and are not stored beyond session needs.
  • No third-party sharing: Your code and metadata are never shared with external parties.

​
Storage and transmission

  • Encrypted at rest and in transit: All data is encrypted using industry-standard protocols (TLS in transit, AES-256 at rest).
  • Isolated indexing: Each organization’s code index is isolated and never mixed with other tenants.

​
Deployment and integration

  • Flexible deployment: Available via cloud, hybrid, or as a Model Context Protocol (MCP) server to support air-gapped or self-managed environments.
  • Audit-ready: All interactions with Qodo Context Engine MCP can be logged and traced, enabling compliance with internal policies and external regulations.

​
Trust and control

  • You own your data: You choose what’s connected, indexed, and queried.
  • Enterprise controls: Role-based access, tagging, and scoping controls help you define how Qodo Context Engine MCP is used across your team.