> ## Documentation Index
> Fetch the complete documentation index at: https://getalchemystai.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Alchemyst MCP Server

> Adding the Alchemyst MCP Server to Github Copilot in VS Code

<iframe width="560" height="315" src="https://www.youtube.com/embed/T0GbQsJszvk" title="Setting up your MCP Server" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen />

## Overview

This tutorial shows how to integrate the **Alchemyst MCP** (Memory & Context Platform) server with **Visual Studio Code** and **GitHub Copilot** to give your AI tooling project-specific memory and documentation. MCP helps your AI suggestions stay aligned with PRDs, coding patterns, and team standards.

## Key Learnings

* What MCP is and why context engineering matters for AI-assisted development.
* How to install and configure an MCP server for VS Code workflows.
* How to upload and link Product Requirement Documents (PRDs).
* How to enable Agent Mode so Copilot receives project context in real time.
* Practical benefits: improved accuracy, faster development, and consistent team-wide behavior.

## Core Concepts

**What is Alchemyst MCP Server?**
A context-engineering platform that supplies project-specific memory and documentation to developer tooling so AI assistants produce suggestions that follow your project's requirements and standards.

**Why it matters**
MCP bridges the gap between generic AI suggestions and project-specific expectations, reducing rework and helping teams move faster with fewer errors.

## Technical Implementation

### Setup Process

1. **Install MCP**:deploy locally for development or containerize (Docker) for team/production use.
2. **Integrate with VS Code**: install/configure the MCP extension or point your extension settings to the MCP endpoint.
3. **Upload PRDs**: add structured project documents (Markdown, PDF, plain text) through the web UI or API and attach metadata (project\_id, version, tags).
4. **Link repositories**: associate PRDs with repositories/workspaces so MCP surfaces the right docs for each codebase.
5. **Activate Agent Mode**: toggle Agent Mode in the extension to pass context into Copilot; set the agent scope (`file`/`module`/`project`) to control influence.

### Configuration Essentials

* Set MCP server URL and `API key` in workspace or extension settings.
* Provide PRD metadata (project\_id, title, version, tags, author, last\_updated).
* Use HTTPS/TLS and RBAC in production.

## Quick Start

1. Deploy MCP (local Docker or hosted).
2. In VS Code, install/configure the Alchemyst MCP extension (set `mcp.server.url` and `mcp.apiKey`).
3. Upload PRDs and tag them by area (e.g., `auth`, `payments`, `frontend`).
4. Link the PRDs to your repository/workspace in the MCP dashboard.
5. Enable **Agent Mode** with a conservative scope (start with `file` or `module`) and test Copilot suggestions.

## Business & Team Impact

* **Increased velocity:** Developers get more relevant suggestions aligned with requirements.
* **Reduced errors:** Fewer mismatches between intended and generated code.
* **Consistency:** Team members and CI/CD checks use the same source of truth.
* **Reliability:** AI assistance becomes dependable for production workflows.

## Practical Applications

* Use PRDs to guide code generation for features (APIs, UX flows, security rules).
* Maintain context across sessions so the AI remembers project constraints.
* Integrate MCP into code review and CI pipelines to surface context-aware guidance.

## Best Practices

* Keep PRDs **concise** and **modular**; split large docs into focused pieces.
* Tag aggressively so MCP can quickly select relevant documents.
* Start Agent Mode with a narrow scope before scaling it to project-wide.
* Version your PRDs and map versions to branches or releases.
* Protect sensitive documents with RBAC and audit logging.

## Troubleshooting

* **Server unreachable:** Check container/process status, open ports, and firewall rules.
* **Auth failures:** Verify API keys and tokens; check server logs for validation errors.
* **Copilot not influenced:** Confirm Agent Mode is enabled and the repo is linked to the correct PRD.
* **Slow responses:** Reduce PRD size or scope; check server resource usage.

## Resources

Explore more materials to continue learning:

* Platform: [Alchemyst AI](https://getalchemystai.com?utm_source=docs.getalchemystai.com\&utm_campaign=docs\&utm_medium=web)
* Documentation: [MCP Server Documentation](https://docs.getalchemystai.com/mcps/introduction)
* Full video playlist: ▶️ [Complete YouTube Playlist](https://www.youtube.com/playlist?list=PLTnYaGyWE7dGlzullvvE_7tXd1cLx9iYg)