Create spaces to organize and centralize relevant content that grounds Copilot’s responses in the right context for a specific task.
Who can use this feature?Anyone with a Copilot license can use Spaces.
Note
Copilot Spaces is in public preview and subject to change.
To create a space, go to https://github.com/copilot/spaces, and click Create space.
Give your space a name.
Choose whether the space is owned by you or by an organization you belong to. Organization-owned Spaces can be shared using GitHub’s built-in permission model.
Optionally, add a description. This does not affect the responses Copilot gives in the space, but it can help others understand the context of the space.
Note
You can change the name and description of your space at any time by clicking Edit in the top right corner of the space.
Click Save in the top right corner of the page.
You can add two types of context to your space:
Instructions: Free text that describes what Copilot should focus on within this space. Include its areas of expertise, what kinds of tasks it should help with, and what it should avoid. This helps Copilot give more relevant responses based on your intent.
For example:
You are a SQL generator. Your job is to take the sample queries and data schemas defined in the attached files and and generate SQL queries based on the user's goals.
Attachments: This context will be used to provide more relevant answers to your questions. Additionally, Spaces will always refer to the latest version of the code on the main
branch of the repository.
To add attachments, click Add to the right of "Attachments", then choose one of the following options:
Once you've added context to your space, you can ask Copilot questions in the chat interface. Your chat will be grounded in the context you've added.
You can also change the large language model (LLM) used for your space by selecting the CURRENT-MODEL dropdown menu, then clicking the AI model of your choice. For more information about choosing the right AI model, see Choosing the right AI model for your task.
Next stepsRetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4