Get Started

Integrating the Pieces MCP with Continue.dev brings your workflow context directly into your editor. Continue.dev is an open-source coding assistant available as a VS Code or JetBrains extension.

**Agent mode required:** MCP servers in Continue.dev only work in **agent mode**. They are not available in regular chat or plan modes.

Prerequisites

There are two prerequisites for integrating Pieces with Continue.dev as an MCP—an active instance of PiecesOS and the fully-enabled Long-Term Memory engine.

Make sure that PiecesOS is installed and running. This is *required* for the MCP server to communicate with your workflow data. Enable the Long-Term Memory Engine (LTM-2.7) through the Pieces Desktop App or the [PiecesOS Quick Menu](/products/core-dependencies/pieces-os/quick-menu) in your toolbar.

Installing PiecesOS & Configuring Permissions

Follow the instructions below for a detailed guide on setting up and configuring PiecesOS to correctly pass captured workflow context to the Pieces MCP server.

Setting Up Continue.dev

Continue.dev uses YAML configuration with an mcpServers array. The recommended location is .continue/config.yaml at your project root (you can commit it to version control for team sharing).

Config File Location

Scope Path
Workspace (recommended) .continue/config.yaml at project root
Alternative JSON .continue/mcpServers/mcp.json

Local Setup (Streamable HTTP — recommended)

Create or edit .continue/config.yaml:

mcpServers:
  - name: Pieces LTM
    type: streamable-http
    url: http://localhost:39300/model_context_protocol/2025-03-26/mcp

Local Setup (SSE)

mcpServers:
  - name: Pieces LTM
    type: sse
    url: http://localhost:39300/model_context_protocol/2024-11-05/sse
Use `streamable-http` (with hyphen), not `streamableHttp` or `http`. The type value is case-sensitive.

Using Pieces MCP Server in Continue.dev

Once integrated, you can utilize Pieces LTM directly in Continue.dev.

Switch Continue.dev to **agent mode** in the UI. MCP servers are only available in agent mode. Check the available tools—Pieces LTM tools should appear. Ask the agent to search your Long-Term Memory. For example: *"What was I working on yesterday?"* Check out this [MCP-specific prompting guide](/products/mcp/prompting) if you want to effectively utilize the Long-Term Memory Engine (LTM-2.7) with your new Pieces MCP server.

Updating

Edit your config.yaml (or JSON file), update the url, and switch to agent mode. Continue picks up the change on the next agent session.

Troubleshooting

If you're experiencing issues integrating Pieces MCP with Continue.dev:

  1. Verify PiecesOS Status: Ensure PiecesOS is actively running on your system.

  2. Use Agent Mode: MCP tools do not appear in regular chat—you must be in agent mode.

  3. YAML Parse Error: Check indentation; YAML is whitespace-sensitive.

  4. Type Value: Use streamable-http (with hyphen), not streamableHttp or http.

  5. Server Not Connecting: Check the PiecesOS port (39300-39333) and URL format.

You're now set to enhance your Continue.dev workflow with powerful context retrieval through Pieces MCP. Happy coding!