Example agents that use Microsoft AutoGen with MultiMail for AI-powered email workflows.

A single AutoGen AssistantAgent that monitors an inbox and auto-responds to customer emails. It reads each email, tags it by category and priority, and drafts a professional reply.

What it demonstrates:

• AutoGen tool calling with the MultiMail Python SDK
• RoundRobinGroupChat with a single agent
• Tagging emails for tracking
• Automated reply generation

A two-agent team using SelectorGroupChat:

• inbox_reader — fetches and presents email content
• triage_agent — classifies emails by category, priority, and sentiment

What it demonstrates:

• Multi-agent coordination with SelectorGroupChat
• Separation of I/O (reading emails) from logic (classification)
• Contact management (auto-adding new senders)
• Structured output (summary table)

1. Install dependencies:

pip install -r requirements.txt

2. Configure environment:

cp .env.example .env

Edit .env with your keys:

• OPENAI_API_KEY — your OpenAI API key (platform.openai.com)
• MULTIMAIL_API_KEY — your MultiMail API key (multimail.dev)
• MULTIMAIL_MAILBOX_ID — your mailbox ID (find it in the MultiMail dashboard)

3. Run an example:

python examples/customer_support_agent.py
python examples/email_triage_agent.py

AutoGen agents use tool calling to interact with MultiMail. Each tool is a plain Python function that calls the MultiMail SDK. AutoGen handles the LLM conversation loop, tool dispatch, and multi-agent coordination.

User Task
 |
 v
AutoGen Agent (GPT-4o)
 |-- decides which tool to call
 v
Tool Function (e.g., check_inbox)
 |-- calls MultiMail Python SDK
 v
MultiMail API (https://api.multimail.dev)
 |-- returns email data
 v
AutoGen Agent
 |-- processes result, calls next tool or responds

MultiMail handles regulatory compliance at the infrastructure layer — no SDK-side code changes needed:

• EU AI Act Article 50: Every AI-sent email includes a cryptographically signed ai_generated disclosure in the X-MultiMail-Identity header
• US State Laws: Maine, New York, California, Illinois — AI disclosure built into email delivery
• CAN-SPAM: Unsubscribe headers and physical address footers on all outbound email
• Formally Verified: Lean 4 proofs of identity header tamper evidence

These examples are current for MultiMail v0.5.4.

See multimail.dev/use-cases/eu-ai-act-email-compliance for details.

MultiMail supports graduated trust levels for agent email access:

Start with gated_send and increase autonomy as you build trust.

• MultiMail documentation
• MultiMail Python SDK
• AutoGen documentation
• AutoGen AgentChat tutorial