GeoBot

A serverless AWS Lambda bot for automated GeoGuessr party management and Slack notifications.

Prerequisites

AWS Credentials Setup

Before you can deploy this project, you need to configure AWS credentials. SST will automatically use your AWS credentials to deploy resources.

Option 1: AWS Credentials File (Recommended)

Create or update your AWS credentials file:

macOS/Linux: ~/.aws/credentials Windows: C:\Users\USER_NAME\.aws\credentials

[default]
aws_access_key_id = YOUR_ACCESS_KEY_ID
aws_secret_access_key = YOUR_SECRET_ACCESS_KEY

If you don't have AWS credentials yet:

1. Create an IAM user
2. Configure the AWS CLI

Option 2: Environment Variables

Set these environment variables:

export AWS_ACCESS_KEY_ID=your_access_key_id
export AWS_SECRET_ACCESS_KEY=your_secret_access_key
# Optional for temporary credentials:
export AWS_SESSION_TOKEN=your_session_token

Multiple AWS Profiles

If you have multiple AWS accounts, you can configure multiple profiles:

[default]
aws_access_key_id = DEFAULT_ACCESS_KEY_ID
aws_secret_access_key = DEFAULT_SECRET_ACCESS_KEY

[staging]
aws_access_key_id = STAGING_ACCESS_KEY_ID
aws_secret_access_key = STAGING_SECRET_ACCESS_KEY

[production]
aws_access_key_id = PRODUCTION_ACCESS_KEY_ID
aws_secret_access_key = PRODUCTION_SECRET_ACCESS_KEY

Then specify the profile in sst.config.ts:

providers: {
  aws: {
    profile: "staging"; // or use input.stage to switch profiles per stage
  }
}

Slack Configuration

You'll need to set up Slack secrets for the bot to work:

# Set your Slack bot token
pnpx sst secret set SlackBotToken "xoxb-your-slack-bot-token"

# Set your Slack channel ID or name
pnpx sst secret set SlackChannel "#your-channel-name"

GitHub Actions (scheduled e2e)

The E2E workflow runs the Playwright test on a schedule against the deployed stage, authenticating to AWS via GitHub OIDC.

The OIDC provider is an account-level singleton that sst.config.ts only references, so it must be created once per AWS account:

aws iam create-open-id-connect-provider \
  --url https://token.actions.githubusercontent.com \
  --client-id-list sts.amazonaws.com \
  --thumbprint-list 6938fd4d98bab03faadb97b34396831e3780aea1

Deploying then creates the role and prints its ARN as the githubE2eRole output. Set these repository variables (Settings → Secrets and variables → Actions → Variables):

• AWS_ROLE_ARN — the githubE2eRole ARN from the deploy output
• AWS_DEPLOY_ROLE_ARN — the githubDeployRole ARN from the deploy output
• AWS_REGION — the region the app is deployed to (must match sst.config.ts)
• SST_STAGE — the stage the workflows deploy and test against

Getting Started

Development

Start the SST development server:

pnpx sst dev

This will:

• Deploy your Lambda functions to AWS
• Set up the cron schedule
• Enable live development with instant updates

Seting up puppeteer for dev

pnpx @puppeteer/browsers install chromium@latest --path /tmp/localChromium

After update your .env YOUR_LOCAL_CHROMIUM_PATH to point to your installed tmp local chromium

Deployment

Deploy to production:

pnpx sst deploy --stage production

Project Structure

• src/create-party.ts - Main Lambda handler for the bot functionality
• sst.config.ts - SST infrastructure configuration with cron scheduling
• sst-env.d.ts - Auto-generated TypeScript definitions for SST resources