Setup Guide

Prerequisites

• Python 3.11+
• An AWS account with Bedrock access (Claude 3.5 Haiku)
• A Microsoft 365 tenant with Teams
• The chukfi CLI binary compiled and on PATH
• A PostgreSQL database for the Chukfi CMS

Step 1: Enable Amazon Bedrock

Request Claude 3.5 Haiku Access

1. Go to the AWS Bedrock console
2. In the left sidebar, click Model access
3. Click Manage model access
4. Find Claude 3.5 Haiku by Anthropic and request access
5. Wait for approval (usually instant for Haiku)

Configure AWS Credentials

The bot uses the standard AWS credential chain (env vars, ~/.aws/credentials, or IAM role):

export AWS_REGION=us-east-1
export AWS_ACCESS_KEY_ID=AKIA...
export AWS_SECRET_ACCESS_KEY=...

For production on ECS, attach an IAM role with the BedrockRuntimeAccess policy.

Verify Bedrock Access

aws bedrock-runtime converse \
  --model-id anthropic.claude-3-5-haiku-20241022-v1:0 \
  --messages '[{"role":"user","content":[{"text":"Hello"}]}]'

Step 2: Set Up the Teams Outgoing Webhook

1. Open Microsoft Teams
2. Go to your team → Apps → Outgoing Webhook
3. Configure:
   • Name: Chukfi CMS Bot
   • Callback URL: https://your-server.com/api/webhook
   • Description: AI-powered content management
4. Click Create
5. Copy the HMAC secret — you’ll need it for the environment config

Step 3: Configure the Bot

Clone and Install

git clone https://github.com/smattera/chukfi-teams-bot.git
cd chukfi-teams-bot
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Environment Variables

Copy the example config and fill in your values:

cp .env.example .env

Variable	Required	Description
AWS_REGION	Yes	AWS region for Bedrock (default: us-east-1)
BEDROCK_MODEL_ID	No	Bedrock model ID (default: anthropic.claude-3-5-haiku-20241022-v1:0)
TEAMS_WEBHOOK_SECRET	Yes	HMAC secret from Teams Outgoing Webhook
CHUKFI_BIN	No	Path to chukfi binary (default: found on PATH)
CHUKFI_CONFIG	No	Path to chukfi.config.json (default: ./chukfi.config.json)
DATABASE_URL	Yes	PostgreSQL connection string
LOG_LEVEL	No	Logging level (default: INFO)

Update Tenant IDs

Edit config.py to set your actual Entra Tenant UUIDs:

ALLOWED_TENANTS: ClassVar[set[str]] = {
    "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee",  # Your tenant UUID
}

Step 4: Run the Bot

Development Mode

# Disable HMAC auth for local testing
TEAMS_WEBHOOK_SECRET="" uvicorn main:app --reload --port 8000

Production Mode

uvicorn main:app --host 0.0.0.0 --port 8000

Test with curl

curl -X POST http://localhost:8000/api/webhook \
  -H "Content-Type: application/json" \
  -d '{"text":"Create a blog post about flu shots","from":{"name":"[email protected]"}}'

Step 5: Deploy

Docker

# Build the image (from the repo root)
docker build -t chukfi-teams-bot -f chukfi-teams-bot/Dockerfile .

# Run with env file
docker run -p 8000:8000 --env-file .env chukfi-teams-bot

ECS Fargate (Recommended)

1. Push the Docker image to Amazon ECR
2. Create an ECS task definition with the image
3. Set environment variables in the task definition or AWS Secrets Manager
4. Attach an IAM task role with Bedrock runtime access
5. Deploy behind an Application Load Balancer
6. Point your domain’s DNS at the ALB

Cloud VM

Any Linux VM works. The bot is lightweight — 0.25 vCPU / 512 MB is sufficient.

Step 6: Test in Teams

1. Go to your team channel in Microsoft Teams
2. Type @Chukfi CMS Bot create a blog post about fall health tips
3. The bot should respond with a confirmation and the entry ID
4. Check the admin UI — the entry should appear in draft status
5. Review and publish via the admin interface