Support

Google Antigravity Setup

Connect Tempreon to Google Antigravity so Gemini-powered agents load your Core Imprint, search your Knowledge Vault, and remember what they learn.

Overview

Google Antigravity is Google's agent-first development environment — an IDE, CLI, and SDK built around Gemini models. It replaced Gemini CLI as the consumer path in June 2026, and it treats MCP servers as first-class: add Tempreon once and every Antigravity agent can use it.

Connecting Tempreon gives those agents persistent intelligence: your AI loads your Core Imprint, searches your Knowledge Vault for context, and remembers what it learns — and because Tempreon is one shared layer, that memory follows you across every other connected tool.

As of July 2026, Antigravity is in free public preview; a Pro tier is included with Google AI Pro, and paid Ultra tiers are available for heavier use. MCP setup works the same way on each.

Quick path: add Tempreon to ~/.gemini/config/mcp_config.json with a serverUrl entry → Agent Settings → Customizations → Authenticate → approve in the browser → copy the authorization code back → Submit.

Prerequisites

  • Google Antigravity installed (IDE or CLI) and signed in with your Google account
  • A Tempreon account with at least one available Bridge slot
  • A browser for the authorization step

Step-by-Step Setup

1. Add Tempreon to the MCP config

Antigravity reads MCP servers from a JSON config file:

  • Global (all projects): ~/.gemini/config/mcp_config.json
  • Workspace (one project): .agents/mcp_config.json in the project root

Add Tempreon using the serverUrl field:

{
  "mcpServers": {
    "tempreon": {
      "serverUrl": "https://api.tempreon.com/functions/v1/tempreon-mcp/mcp"
    }
  }
}

The field name matters. Antigravity accepts serverUrl only — it rejects configs that use url or httpUrl. If the server never shows up, this is the first thing to check. (If you're migrating a config from Gemini CLI, which uses httpUrl, rename the field.)

The URL is also shown in your Bridges dashboard. You don't need a token, API key, or client ID — Tempreon supports dynamic client registration, so OAuth is zero-config.

2. Authenticate

  1. In Antigravity, open Agent SettingsCustomizations
  2. Find the tempreon server and select Authenticate
  3. A browser tab opens to Tempreon's consent screen — sign in to your Tempreon account and approve
  4. The browser then shows an authorization code. Copy it — this is a paste-back flow, not an automatic redirect. If you close the tab without copying, restart from Authenticate
  5. Return to Antigravity, paste the code, and select Submit

Tokens are stored at ~/.gemini/antigravity/mcp_oauth_tokens.json.

3. Verify the connection

In the Antigravity CLI, run /mcp to check server status. In the IDE, ask an agent "What Tempreon tools do you have access to?" — if it answers with a list, the Bridge is live.

Troubleshooting

The server never appears, or the config is rejected. Check the field name: it must be serverUrl. Antigravity rejects url and httpUrl.

Connection fails immediately. Confirm the URL ends in /mcp with no trailing slash: https://api.tempreon.com/functions/v1/tempreon-mcp/mcp.

You approved in the browser but Antigravity still shows the server as unauthenticated. The flow isn't done at approval — copy the authorization code from the browser, paste it back in Antigravity, and select Submit.

Authentication keeps failing. Run the Authenticate flow again from Agent Settings → Customizations. Stored tokens live at ~/.gemini/antigravity/mcp_oauth_tokens.json.

Coming from Gemini CLI? Google stopped serving Gemini CLI for consumer accounts on June 18, 2026 — Antigravity is the replacement. The config formats differ (serverUrl here, httpUrl there), so re-create the entry rather than copying it across.

Also connect: Gemini App · Gemini CLI · Any MCP Client

Try saying...

  • "Start my session." — triggers session_start to load your Core Imprint
  • "Search my knowledge base for [topic] before we plan this change." — pulls your Knowledge Vault into the agent's context
  • "Remember that [decision] — we chose it because [reason]." — stores the insight for every future session, in every connected tool