Build an AI Email Writer: Draft Professional Emails in Seconds

Project Overview

An AI tool that turns brief descriptions, bullet points, or rough drafts into polished professional emails. Supports tone selection, email type templates, and length control.

Learning Goals

• Prompt engineering for consistent tone and style
• Build a Streamlit form-based UI
• Use temperature settings to control creativity
• Handle different email types with templates

Implementation

Step 1: Setup

pip install openai streamlit

Step 2: Email Writer Core

# email_writer.py
from openai import OpenAI

client = OpenAI()

EMAIL_PROMPT = """Write a professional email based on the following brief.

Email Type: {email_type}
Tone: {tone}
Key Points to Include:
{key_points}

Additional Context: {context}

Requirements:
- Write a complete email with subject line, greeting, body, and sign-off
- Tone must be: {tone}
- Length: {length}
- Be specific and actionable
- Subject line should be clear and compelling

Format the output as:
Subject: [subject line]

[Full email body]"""

TONES = ["Professional", "Friendly", "Formal", "Persuasive", "Empathetic", "Concise"]
EMAIL_TYPES = [
    "Job Application",
    "Follow-up",
    "Meeting Request",
    "Project Update",
    "Customer Support",
    "Sales Outreach",
    "Networking",
    "Apology / Issue Resolution",
    "Thank You",
    "Introduction",
]
LENGTHS = ["Brief (3-4 sentences)", "Standard (2-3 paragraphs)", "Detailed (4+ paragraphs)"]


def write_email(
    key_points: str,
    email_type: str = "Professional",
    tone: str = "Professional",
    context: str = "",
    length: str = "Standard (2-3 paragraphs)",
) -> str:
    response = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[
            {"role": "user", "content": EMAIL_PROMPT.format(
                email_type=email_type,
                tone=tone,
                key_points=key_points,
                context=context,
                length=length,
            )}
        ],
        max_tokens=800,
        temperature=0.7,
    )
    return response.choices[0].message.content


def improve_email(original_email: str, improvement_notes: str) -> str:
    """Improve an existing email draft."""
    response = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[
            {"role": "user", "content": f"""Improve this email draft based on the feedback.

Original Email:
{original_email}

Improvement Notes: {improvement_notes}

Return the improved email with subject line."""}
        ],
        max_tokens=800,
        temperature=0.5,
    )
    return response.choices[0].message.content

Step 3: Streamlit App

# app.py
import streamlit as st
from email_writer import write_email, improve_email, TONES, EMAIL_TYPES, LENGTHS

st.set_page_config(page_title="AI Email Writer", page_icon="✉️")
st.title("✉️ AI Email Writer")
st.caption("Turn bullet points into polished professional emails")

with st.form("email_form"):
    col1, col2 = st.columns(2)

    with col1:
        email_type = st.selectbox("Email Type", EMAIL_TYPES)
        tone = st.selectbox("Tone", TONES)
        length = st.selectbox("Length", LENGTHS)

    with col2:
        context = st.text_area("Context (optional)",
            placeholder="Who are you writing to? What's the background?",
            height=100)

    key_points = st.text_area(
        "Key Points to Include *",
        placeholder="• Ask for a meeting next week\n• Mention our previous conversation\n• Emphasize my experience with Python",
        height=150,
    )

    submitted = st.form_submit_button("Generate Email ✉️", type="primary", use_container_width=True)

if submitted:
    if not key_points.strip():
        st.error("Please enter at least one key point.")
        st.stop()

    with st.spinner("Writing your email..."):
        email = write_email(key_points, email_type, tone, context, length)

    st.divider()
    st.subheader("Generated Email")

    # Show in editable text area
    edited_email = st.text_area("Edit if needed:", value=email, height=300)

    col_a, col_b, col_c = st.columns(3)
    with col_a:
        st.download_button("Download .txt", edited_email, "email.txt", "text/plain")
    with col_b:
        if st.button("Copy to Clipboard"):
            st.write("✅ Copied!")
    with col_c:
        improve_notes = st.text_input("Improve it:", placeholder="Make it shorter and more direct")

    if improve_notes and st.button("Apply Improvement"):
        with st.spinner("Improving..."):
            improved = improve_email(edited_email, improve_notes)
        st.text_area("Improved Email:", value=improved, height=300)

Step 4: Run

streamlit run app.py

Extension Ideas

1. Gmail integration — connect to Gmail API to send directly
2. Reply mode — paste an incoming email and generate a reply
3. A/B testing — generate 2-3 variants and pick the best
4. Language support — add language selector for non-English emails
5. Email thread summarizer — summarize long email threads

What to Learn Next

• AI agents with email tools → AI Agent Fundamentals
• Building a full chatbot → Build an AI Chatbot