Getting started

This guide gets you from signup to your first useful chat in CoopAI.

What you'll need

• VS Code 1.85 or later
• A CoopAI account (free signup or invited by your org admin)
• A local workspace folder open in VS Code

Step 1 — Create an account

Browser → coop-ai.dev/signup/free

Enter your work email and create a password, or continue with Google. Coop sends a verification link on password signup; you can sign in before verifying (soft verify).

For Pro or Enterprise, start at Pricing and follow checkout to the Welcome page. Sign in with the same email you used at checkout.

Forgot your password? → coop-ai.dev/forgot-password

Step 2 — Install the extension

See Install the VS Code extension for marketplace and manual install options.

Step 3 — Sign in

1. Extension UI — Open Coop sidebar → gear icon (or CoopAI: Open Settings).
2. Account — Choose a sign-in path:

Path	Steps
Continue with Google	Click the Google button at the top
Continue with email	Enter email → Continue with email → password → Sign in
Sign in with SSO	Enterprise: enter Organization name → Sign in with SSO → finish in your browser

3. Success: Account shows your org and plan. Open chat and ask a question.

Enterprise SSO: Organization name is required before Sign in with SSO. Coop opens your system browser for IdP login; VS Code completes the session when you return. Name matching is case-insensitive.

Automation API keys are for CI/scripts only — create them in the admin portal if needed, not in the extension.

Step 4 — Ask your first question

Open the Coop icon in the activity bar. Click a quick action or type in the composer.

1. Open a file in your workspace.
2. Extension UI — Type in the chat composer or click Understand Repo.
3. Try: Explain this codebase. What are the main entry points?

Coop streams an answer grounded in your workspace files and any connected integrations. Plain chat uses GPT-4o mini — see Model assignments.

Step 5 — Try a quick action

With a file open, highlight a few lines of code, then right-click the selection. Choose CoopAI: Trace Decision for Selection (or Find Owner, Blast Radius, Understand Repo, Knowledge Gaps).

You can also type /trace in the chat composer for the same action.

Quick actions route to Claude Sonnet 4.6 for structured, repo-grounded answers.

Inline autocomplete

CoopAI shows ghost-text code completions by default as you type.

The sidebar header shows Autocomplete On for most installs. To turn it off: click Autocomplete in the header, or open CoopAI Settings → Preferences → Model & chat and uncheck Enable inline autocomplete → Save model settings.

Type in a code file — ghost text appears after a short pause; Tab to accept.

See Inline autocomplete for FIM, graph context (indexed repos), Copilot coexistence, and how to turn off intentionally.

Verify production (smoke check)

After sign-in, confirm the hot path:

Check	Success looks like
API health	Browser → api.coop-ai.dev/health returns OK
Model & chat	Settings → Preferences → Model & chat — four assigned models (Chat, Quick actions, /edit, Autocomplete), no provider picker
Autocomplete	Sidebar Autocomplete On → type in .ts → ghost text → Tab accepts
Edit mode	/edit add a comment above this function → Apply / Undo on the patch notification
Quick action	Understand Repo or /trace on a selection → structured answer with sources
Plain chat	Composer question → grounded reply (GPT-4o mini)

More fixes: Troubleshooting.

Plans at a glance

Plan	Best for
Developer (free)	Individual use — full integrations, 3-repo Deep-Index cap, solo seat
Pro	Teams — unlimited indexing, team seats ($20/seat/month), usage analytics
Enterprise	Self-hosted, BYOK, zero-retention, SAML SSO, compliance

Enterprise org admins: configure SAML in the admin portal before rolling SSO to your team — see Single Sign On (SSO).

See Plans & billing for details.

Next

• Model assignments — per-feature models and settings UI
• Inline autocomplete — ghost-text completions
• Extension settings — workspace, tools, preferences
• Owner's Manual — daily use, quick actions, prompt library