Skip to main content
Hyper’s shared memory layer is most powerful when your whole team is connected to the same workspace. When everyone’s AI tools draw from the same pool of memories — architecture decisions, coding conventions, resolved bugs, project goals — you stop re-explaining context and start compounding knowledge. This guide walks you through creating a workspace, inviting members, and getting everyone set up for day one.

​
Step 1: Create Your Workspace

Your workspace is the shared memory container for your team. One person (the admin) creates it during the Hyper desktop app’s onboarding flow.
1

Download and open the Hyper desktop app

If you haven’t installed Hyper yet, download the desktop app from gethyper.space and launch it.
2

Sign in with Google

Authenticate with your Google account. This becomes your Hyper identity across all AI tools.
3

Create a workspace

During onboarding, choose Create a new workspace when prompted. Give it a name that reflects your team or project — for example, acme-eng or product-team.
4

Complete setup

Hyper will initialize your workspace with a default structure. You’re now the admin of this workspace.
The person who creates the workspace becomes the admin and has full control over members, integrations, and settings. Choose someone who will actively manage the workspace long-term.

​
Step 2: Invite Team Members

Once your workspace exists, generate an invite link to share with your team.
1

Open Workspace Settings

In the Hyper desktop app, go to Settings → Workspace.
2

Generate an invite link

Click Invite. Hyper generates a unique join_token and wraps it into a shareable link.
3

Share the link

Send the link to teammates via Slack, email, or any channel your team uses. Each link is tied to your workspace and can be reused for multiple people.
Invite links contain a join_token that grants access to your workspace. Treat them like a password — share only with people you intend to add and rotate the link if it’s accidentally exposed.

​
Step 3: Join a Workspace

Team members can join in two ways — through the desktop app UI, or directly via their AI agent if they already have Hyper connected.
1

Open the Hyper desktop app

Download and install the desktop app if needed, then launch it.
2

Sign in with Google

Authenticate with your Google account.
3

Join the workspace

When prompted during onboarding, choose Join a workspace. Paste the invite link (or the raw join_token) that the admin sent you and confirm.
4

Connect your AI tools

Once you’ve joined, follow the Claude setup guide or Cursor setup guide to connect Hyper to your AI tools.

​
Step 4: Set Up Team Integrations

After everyone has joined, connect the tools your team already uses so Hyper can pull context from them automatically.
1

Open Connections settings

In the Hyper desktop app, go to Settings → Connections.
2

Connect your integrations

Add the services your team relies on — for example, GitHub, Linear, Notion, or Slack. Each integration allows Hyper to observe and remember context from that source.
3

Encourage each member to connect

Integrations are per-user. Ask every team member to visit Settings → Connections after joining and connect their own accounts.
Workspace admins can see which integrations members have connected, but they cannot access the underlying credentials. Each member authorizes their own services independently.
The shared workspace becomes genuinely powerful once you seed it with the foundational context your team returns to again and again. We recommend creating three files in your workspace’s org/ folder during your first team session.

org/identity.md

Describe who your team is: your mission, tech stack, team structure, and the conventions your AI tools should always respect. This file loads every time someone calls connect.

org/decisions.md

A running log of significant architectural and product decisions — what you chose, what you rejected, and why. Prevents relitigating the same debates.

org/goals.md

Your current sprint goals, quarterly OKRs, or product roadmap. Gives AI tools the “why” behind requests, leading to more relevant suggestions.

Regular review

Schedule a short monthly review to keep these files current. Stale context is worse than no context — AI tools will confidently surface outdated information.

​
Sample org/identity.md

Here’s a starting template you can paste into your workspace and customize:
org/identity.md
# Team Identity

## Who We Are
[Your team name] builds [product description] for [target users].

## Tech Stack
- Backend: [e.g., Python / FastAPI]
- Frontend: [e.g., React / TypeScript]
- Infrastructure: [e.g., AWS / Terraform]
- Primary language: [e.g., English]

## Conventions
- All PRs require two approvals before merge
- We prefer explicit error handling over silent failures
- API endpoints follow REST conventions unless documented otherwise

## Team Members
- [Name] — [Role]
- [Name] — [Role]
Have each team member add a personal section to org/identity.md describing their role and areas of ownership. When someone asks the AI a question, it can route context to the right person or surface memories from the right part of the codebase.

​
Roles at a Glance

RoleCapabilities
AdminFull access: invite members, manage integrations, edit workspace settings, read and write all workspace files
MemberWrite to their own people/{name}/* files, their assigned areas, and feed.md; read shared workspace memories and use all MCP tools