> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pre.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# OAuth Connectors

> Sign in once to let pre.dev read and write to external services on your behalf.

OAuth Connectors let pre.dev act on your behalf in external services — reading data, creating resources, or keeping state in sync — without you pasting API keys. You click "Connect," sign in once, and the agent inherits that authorization for every project.

## How it works

1. Open **[Integrations → OAuth Connectors](https://pre.dev/projects/integrations)**
2. Pick a service and click **Connect**
3. Sign in through the service's standard OAuth flow — the service shows you a consent screen listing exactly what access pre.dev is requesting, and nothing is granted until you approve
4. Done — any project the agent builds can now use that connector

## Available connectors

**Project management**

* **Linear** — create and update issues; sync your roadmap into a Linear project
* **Jira** (Atlassian) — create and update tickets; sync your roadmap into a Jira project
* **Asana** — tasks and projects
* **Monday.com** — boards and items

**Communication**

* **Slack** — send messages, post updates to channels
* **Discord** — interact with your servers

**Docs, files & knowledge**

* **Notion** — read briefs and docs, write pages and databases
* **Google Workspace** — Sheets, Drive, Docs, Calendar, Gmail
* **Microsoft** — Microsoft Graph (Outlook, OneDrive, Teams data)
* **Dropbox** — read and write files

**Design**

* **Figma** — read design files and components

**Data & CRM**

* **Airtable** — read and write bases
* **HubSpot** — contacts, deals, CRM records
* **Salesforce** — objects and records in your org

**Source control**

* **GitHub** — repositories for your project's code (see [Pull Requests](/coding-agent/building/pull-requests))
* **GitLab** — API access to your GitLab resources
* **Bitbucket** — repository access

## What agents do with them

When a build task mentions a connected service, the agent uses the connector directly:

* **"Create a Linear issue for every failing test"** → agent uses your Linear connector
* **"Send a Slack message when a PR merges"** → agent uses your Slack connector
* **"Read my Notion project brief before starting"** → agent reads via your Notion connector
* **"Pull the pricing table from my Google Sheet"** → agent reads via your Google connector
* **"Log new signups as HubSpot contacts"** → agent writes via your HubSpot connector

If the agent needs a service you haven't connected yet, it tells you and points you to the Integrations page.

The agent never sees your password. It holds a scoped token that you can revoke from the Integrations page at any time.

## Security

* Tokens stay server-side — the agent calls each service through pre.dev's integration layer, so tokens are never embedded in your project's code or exposed to the browser
* Scopes are shown on the consent screen before you authorize — pre.dev requests the minimum required
* Expiring tokens are refreshed automatically; if a refresh fails, the agent asks you to reconnect rather than failing silently
* You can disconnect any connector from the Integrations page and all active tokens are revoked