Skip to content

Email and Calendar management via browser automation (Playwright) #660

@kovtcharov

Description

@kovtcharov

Summary

Enable GAIA to manage email (Gmail, Outlook) and calendar (Google Calendar, Outlook Calendar) through Playwright browser automation. This is the fastest path to Tier 1 use-case enablement — no dedicated API integrations needed.

Strategic Context

From the OpenClaw strategy (§9.5):

Email triage and management — Always-on: Yes, Sensitive Data: Yes, AMD Local Advantage: Very strong (cost + privacy) — Tier 1: Lead with

Why browser automation first:

  • Works with ANY web app — Gmail, Outlook, Google Calendar, etc.
  • No API keys or OAuth setup required — user just logs in
  • Sandboxed via Playwright — safe, cross-platform
  • Enables demos immediately — "tell my agent to check my email" works day one
  • Build dedicated API integrations later where performance matters

Use Cases Enabled

Email (via Gmail/Outlook web)

  • Read and summarize unread emails
  • Draft and send replies
  • Search for specific emails
  • Archive/label/organize messages
  • "Check my email and summarize anything urgent"

Calendar (via Google Calendar/Outlook web)

  • View upcoming events and daily agenda
  • Create new events from natural language
  • Check availability for meeting scheduling
  • "What's on my calendar today?" / "Schedule a meeting with John at 3pm Friday"

General Web App Workflows

  • Fill out forms on any website
  • Extract data from dashboards
  • Monitor web pages for changes
  • "Go to my banking dashboard and show me this month's spending"

Architecture

Built on #458 (BrowserToolsMixin). This issue adds:

  1. Pre-built workflow templates for common email/calendar actions
  2. Auth persistence — Playwright browser context saves login sessions across agent restarts
  3. Structured extraction — Parse email lists, calendar events into structured data the agent can reason about

Dependencies

Acceptance Criteria

  • Agent can log into Gmail and summarize unread emails
  • Agent can draft and send an email reply
  • Agent can check Google Calendar and list today's events
  • Agent can create a calendar event from natural language
  • Browser login sessions persist across agent restarts
  • Works in headless mode (no visible browser window)
  • Demo: "Check my email and tell me what's urgent, then look at my calendar for tomorrow"

Metadata

Metadata

Assignees

No one assigned

    Labels

    browser-useBrowser automation and control featuresdomain:agent-coreFramework, tools, registry, memory, skills, orchestrationenhancementNew feature or requesttrack:consumer-appHermes-competitor consumer product — mobile-first, voice + messaging + memory + skills

    Type

    No type

    Projects

    No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions