Pieces + Github Copilot

Access your Pieces Long-Term Memory from inside GitHub Copilot inside your IDE of choice, powered by the Pieces MCP server. Bring your Long-Term Memory into your agent chats, allowing GitHub Copilot to suggest and make code changes based on the context captured across all your activities in your developer workflow.

Use case

Github Copilot

Github Copilot + Pieces

🧠 Available context

Current IDE only

Current IDE + all your activities

🧠 Available context

Current IDE only

Current IDE + all your activities

✅ Access to GitHub issues, Jira tickets, or other work tracking in chat

Not supported

Access any issue or ticket you have looked at across any repo

✅ Access to GitHub issues, Jira tickets, or other work tracking in chat

Not supported

Access any issue or ticket you have looked at across any repo

🛠️ AI generated commit messages

Only from code changes

From code changes and any other relevant context like conversations with colleagues

🛠️ AI generated commit messages

Only from code changes

From code changes and any other relevant context like conversations with colleagues

🤖 Agentic code changes based on conversations with colleagues

Manual by copy/paste instructions

Ask Pieces to retrieve conversations, and have GitHub Copilot update the code

🤖 Agentic code changes based on conversations with colleagues

Manual by copy/paste instructions

Ask Pieces to retrieve conversations, and have GitHub Copilot update the code

✍️ AI-assisted code documentation

Limited to open project only

Generate documentation using the open project, and other context from documents, conversations, and more

✍️ AI-assisted code documentation

Limited to open project only

Generate documentation using the open project, and other context from documents, conversations, and more

🧠 Access to how you solved bugs or errors in the past

None

Pieces can recall the last time you hit an error and what you did

🧠 Access to how you solved bugs or errors in the past

None

Pieces can recall the last time you hit an error and what you did

Access your Long-Term Memories from the GitHub Copilot agent chat

Connect the Pieces MCP server to GitHub Copilot, allowing your Long-Term memory to be queried from inside the agent chat. Mix your day to day coding prompts with asking Pieces questions based on your activities.

Prompt by time range or source application

Pieces inside GitHub Copilot supports time-based and source-based queries, allowing you to ask questions like “What was I doing yesterday?” or “Summarize all my emails from the last week in Superhuman”.

Use the responses from Pieces to update code

Use the responses from Pieces into your GitHub Copilot agentic workflows. Ask GitHub Copilot to retrieve details of a code change you were discussing with a colleague from the Pieces Long-Term memory, and have the agent implement the changes automatically.

01

Step 1

Get the MCP URL from the Pieces menu bar. Pieces supports MCP over SSE.

01

Step 1

Get the MCP URL from the Pieces menu bar. Pieces supports MCP over SSE.

01

Step 1

Get the MCP URL from the Pieces menu bar. Pieces supports MCP over SSE.

02

Step 2

Inside the GitHub Copilot configuration in your IDE of choice, add a new MCP server, passing in the Pieces MCP server URL. Check out our guide for configuring Pieces in GitHub Copilot.

02

Step 2

Inside the GitHub Copilot configuration in your IDE of choice, add a new MCP server, passing in the Pieces MCP server URL. Check out our guide for configuring Pieces in GitHub Copilot.

02

Step 2

Inside the GitHub Copilot configuration in your IDE of choice, add a new MCP server, passing in the Pieces MCP server URL. Check out our guide for configuring Pieces in GitHub Copilot.

03

Step 3

Ask Pieces a question using the GitHub Copilot agent chat. Pieces will extract the relevant context from the Long-Term Memory and return it to GitHub Copilot for processing. Check out our MCP prompting guide for more details.

03

Step 3

Ask Pieces a question using the GitHub Copilot agent chat. Pieces will extract the relevant context from the Long-Term Memory and return it to GitHub Copilot for processing. Check out our MCP prompting guide for more details.

03

Step 3

Ask Pieces a question using the GitHub Copilot agent chat. Pieces will extract the relevant context from the Long-Term Memory and return it to GitHub Copilot for processing. Check out our MCP prompting guide for more details.

Frequently asked questions

How do I connect the Pieces MCP server to GitHub Copilot?

Pieces supports the SSE transport for MCP, and you can get the URL from the PiecesOS menu bar. It will be something like http://localhost:39300/model_context_protocol/2024-11-05/sse, but the port may change to avoid clashing with other locally running applications. Check the guide on configuring GitHub Copilot in our documentation.

What does the Pieces GitHub Copilot integration actually do?

What LLM does Pieces use to return a response to the request from GitHub Copilot?

Does the Pieces share all my memories to GitHub Copilot by default?

What chat mode is Pieces available in?