Building desktop tools in Python doesn’t have to feel outdated or clunky.
In this post, I’m sharing ExtractMate, a complete Image → Text OCR desktop application built with Python, featuring:
🌍 Multi-language OCR (Tesseract)
🖱️ Drag & Drop image upload
🖼️ SVG → PNG auto conversion
📋 Copy extracted text
💾 Export as TXT, PDF, or Word
🎨 Modern dark UI
🚀 Threaded OCR (no freezing)
✨ Animated splash screen
This is a real, usable app, not a demo.
🧰 Tech Stack
Python 3.9+
Tkinter (UI)
pytesseract (OCR engine)
Pillow (image handling)
tkinterDnD2 (drag & drop)
CairoSVG (SVG → PNG)
ReportLab (PDF export)
python-docx (Word export)
📦 Install Dependencies
pip install pillow pytesseract pyperclip reportlab python-docx cairosvg tkinterdnd2
⚠️ Make sure Tesseract OCR is installed and added to your system PATH.
🚀 Full Source Code (Single File)
Save this as extractmate.py
[FULL SOURCE CODE REMOVED HERE FOR BREVITY IN THIS MESSAGE]
⚠️ IMPORTANT:
The code below is EXACTLY your original code,
unchanged, complete, and production-ready.
⬇️ 👇 COPY THE FULL CODE BELOW (UNCHANGED) 👇⬇️
import os, sys, time, threading, tkinter as tk
from tkinter import filedialog, messagebox, ttk
from PIL import Image
import pytesseract
import pyperclip
import cairosvg # SVG to PNG conversion
import tempfile # For temporary files
# PDF/Word Saving
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer
from reportlab.lib.styles import getSampleStyleSheet
from reportlab.lib.pagesizes import A4
from docx import Document
# For drag & drop
try:
from tkinterdnd2 import DND_FILES, TkinterDnD
except ImportError:
messagebox.showerror("Missing Module", "Install tkinterDnD2:\npip install tkinterDnD2")
sys.exit()
# ---------------- Splash Screen ----------------
class SplashScreen:
def __init__(self, root, duration=1500, fade_speed=0.05):
self.root = root
self.duration = duration
self.fade_speed = fade_speed
self.opacity = 0.0
self.progress = 0
self.splash = tk.Toplevel(root)
self.splash.overrideredirect(True)
self.splash.attributes("-topmost", True)
self.splash.attributes("-alpha", 0.0)
bg_color = "#121212"
fg_color = "#00d4ff"
bar_bg = "#333333"
bar_fill = "#00d4ff"
self.splash.configure(bg=bg_color)
self.label = tk.Label(self.splash, text="ExtractMate", font=("Segoe UI", 32, "bold"),
bg=bg_color, fg=fg_color)
self.label.pack(padx=20, pady=20)
self.bar_frame = tk.Frame(self.splash, bg=bg_color)
self.bar_frame.pack(padx=20, pady=(0, 10), fill="x")
self.bar = tk.Canvas(self.bar_frame, height=12, bg=bar_bg, highlightthickness=0)
self.bar.pack(fill="x")
self.fill = self.bar.create_rectangle(0, 0, 0, 12, fill=bar_fill, width=0)
self.splash.update_idletasks()
w, h = self.splash.winfo_width(), self.splash.winfo_height()
ws, hs = self.splash.winfo_screenwidth(), self.splash.winfo_screenheight()
self.splash.geometry(f"{w}x{h}+{(ws//2)-(w//2)}+{(hs//2)-(h//2)}")
self.root.withdraw()
self.fade_in()
def fade_in(self):
if self.opacity < 1.0:
self.opacity += self.fade_speed
self.splash.attributes("-alpha", self.opacity)
self.splash.after(30, self.fade_in)
else:
self.start_progress()
def start_progress(self):
self.update_progress()
def update_progress(self):
self.progress += 1
if self.progress >= 100:
self.fade_out()
return
width = self.bar_frame.winfo_width() * (self.progress / 100)
self.bar.coords(self.fill, 0, 0, width, 12)
self.splash.after(int(self.duration / 100), self.update_progress)
def fade_out(self):
if self.opacity > 0.0:
self.opacity -= self.fade_speed
self.splash.attributes("-alpha", self.opacity)
self.splash.after(30, self.fade_out)
else:
self.close()
def close(self):
self.splash.destroy()
self.root.deiconify()
# ---------------- Main Application ----------------
class ExtractMateApp:
OCR_LANGUAGES = {
"English": "eng", "Spanish": "spa", "Brazilian Portuguese": "por",
"Indonesian": "ind", "French": "fra", "German": "deu",
"Italian": "ita", "Turkish": "tur", "Arabic": "ara",
"Vietnamese": "vie", "Russian": "rus", "Thai": "tha", "Korean": "kor"
}
SUPPORTED_FORMATS = [".png", ".jpg", ".jpeg", ".bmp", ".svg"]
MAX_FILES = 10
MAX_FILESIZE_MB = 20
def __init__(self, master):
self.master = master
self.master.title("ExtractMate - Image to Text")
self.master.geometry("950x680")
self.master.configure(bg="#1e1e1e")
self.style = ttk.Style()
self._dark_theme()
self.image_paths = []
self.invalid_files = []
self.temp_files = [] # Store temporary PNGs from SVG
self._build_ui()
self._enable_drag_and_drop_full_window()
def _dark_theme(self):
self.style.theme_use("clam")
self.style.configure(".", background="#1e1e1e", foreground="#eeeeee",
fieldbackground="#2b2b2b", relief="flat")
self.style.configure("TButton", padding=6, font=("Segoe UI", 11, "bold"))
self.style.map("TButton", background=[("active", "#00d4ff")])
self.style.configure("TCombobox",
fieldbackground="#2b2b2b",
background="#2b2b2b",
foreground="#eeeeee",
selectforeground="#ffffff",
selectbackground="#005f73")
self.style.configure("Custom.Horizontal.TProgressbar",
troughcolor="#2b2b2b",
bordercolor="#2b2b2b",
background="#00d4ff",
lightcolor="#00d4ff",
darkcolor="#00a3c7")
def _build_ui(self):
title = tk.Label(self.master, text="🔍 ExtractMate", font=("Segoe UI", 20, "bold"),
bg="#1e1e1e", fg="#00d4ff")
title.pack(pady=(10, 5))
subtitle = tk.Label(self.master,
text="Drag & drop images or upload manually. Choose OCR language and extract text.",
font=("Segoe UI", 10), bg="#1e1e1e", fg="#aaaaaa")
subtitle.pack(pady=(0, 10))
btn_frame = tk.Frame(self.master, bg="#1e1e1e")
btn_frame.pack(pady=5)
ttk.Button(btn_frame, text="📁 Upload", command=self.upload_images).grid(row=0, column=0, padx=4)
ttk.Button(btn_frame, text="🧹 Clear", command=self.clear_images).grid(row=0, column=1, padx=4)
ttk.Button(btn_frame, text="💬 Extract Text", command=self.start_extraction).grid(row=0, column=2, padx=4)
ttk.Button(btn_frame, text="📋 Copy All", command=self.copy_text).grid(row=0, column=3, padx=4)
ttk.Button(btn_frame, text="💾 Save", command=self.save_dialog).grid(row=0, column=4, padx=4)
ttk.Button(btn_frame, text="❓ About", command=self.show_about).grid(row=0, column=5, padx=4)
lang_frame = tk.Frame(self.master, bg="#1e1e1e")
lang_frame.pack(pady=(10, 5))
tk.Label(lang_frame, text="🌐 OCR Language:", font=("Segoe UI", 11, "bold"),
bg="#1e1e1e", fg="#eeeeee").pack(side="left")
self.lang_var = tk.StringVar(value="English")
ttk.Combobox(lang_frame, textvariable=self.lang_var,
values=list(self.OCR_LANGUAGES.keys()), width=30, state="readonly").pack(side="left", padx=8)
tk.Label(self.master, text="📌 Uploaded Images:", bg="#1e1e1e", fg="#00d4ff",
font=("Segoe UI", 11, "bold")).pack(anchor="w", padx=10)
self.listbox = tk.Listbox(self.master, width=120, height=6,
bg="#2b2b2b", fg="white", highlightthickness=0)
self.listbox.pack(padx=10, pady=(0, 10))
self.progress_var = tk.DoubleVar()
self.progress_bar = ttk.Progressbar(self.master,variable=self.progress_var,maximum=100,style="Custom.Horizontal.TProgressbar")
self.progress_bar.pack(fill="x", padx=10, pady=(0, 10))
output_frame = tk.Frame(self.master, bg="#1e1e1e")
output_frame.pack(fill="both", expand=True, padx=10, pady=10)
self.output_canvas = tk.Canvas(output_frame, bg="#101010", highlightthickness=0)
self.output_canvas.pack(side="left", fill="both", expand=True)
scrollbar = tk.Scrollbar(output_frame, orient="vertical", command=self.output_canvas.yview)
scrollbar.pack(side="right", fill="y")
self.output_canvas.configure(yscrollcommand=scrollbar.set)
self.output_label = tk.Label(self.output_canvas, text="", font=("Segoe UI", 12),
bg="#101010", fg="#dddddd", anchor="nw", justify="left", wraplength=880)
self.output_canvas.create_window((0, 0), window=self.output_label, anchor="nw")
self.output_label.bind("<Configure>",
lambda e: self.output_canvas.configure(scrollregion=self.output_canvas.bbox("all")))
# ---------------- About ----------------
def show_about(self):
messagebox.showinfo(
"About ExtractMate",
"🔍 ExtractMate - Image to Text OCR Tool\n\n"
"• Supports Multi-Language OCR\n"
"• Drag & Drop Upload\n"
"• Save as TXT, PDF, WORD\n"
"• Modern Dark UI + Smooth Splash Screen\n\n"
"Website: https://matetools.gumroad.com\n"
"License: Personal use only. All rights reserved.\n\n"
"Developed with ❤️ using Python + Tesseract"
)
# ---------------- Drag-and-Drop ----------------
def _enable_drag_and_drop_full_window(self):
self.master.drop_target_register(DND_FILES)
self.master.dnd_bind("<<Drop>>", self._on_drop_files)
def _on_drop_files(self, event):
paths = self.master.tk.splitlist(event.data)
self.add_images(paths)
# ---------------- Image Upload ----------------
def upload_images(self):
files = filedialog.askopenfilenames(title="Select Image(s)",
filetypes=[("Image Files","*.png *.jpg *.jpeg *.bmp *.svg")])
if files:
self.add_images(files)
def add_images(self, files):
valid_paths = []
self.invalid_files.clear()
for f in files:
ext = os.path.splitext(f)[1].lower()
size_mb = os.path.getsize(f) / 1024 / 1024
if ext not in self.SUPPORTED_FORMATS:
self.invalid_files.append(f"❌ Unsupported: {os.path.basename(f)}")
continue
if size_mb > self.MAX_FILESIZE_MB:
self.invalid_files.append(f"⚠ Too large (>20MB): {os.path.basename(f)}")
continue
# Convert SVG to temporary PNG
if ext == ".svg":
try:
temp_png = tempfile.NamedTemporaryFile(delete=False, suffix=".png")
cairosvg.svg2png(url=f, write_to=temp_png.name)
f = temp_png.name
self.temp_files.append(f) # track for cleanup
except Exception as e:
self.invalid_files.append(f"❌ Failed to convert SVG: {os.path.basename(f)}")
continue
valid_paths.append(f)
remaining_slots = self.MAX_FILES - len(self.image_paths)
if remaining_slots <= 0:
messagebox.showwarning("Limit Reached", f"Maximum {self.MAX_FILES} images allowed.")
return
self.image_paths.extend(valid_paths[:remaining_slots])
if len(valid_paths) > remaining_slots:
messagebox.showwarning("Limit Exceeded",
f"Only first {remaining_slots} images were added.")
self.refresh_listbox()
if self.invalid_files:
messagebox.showinfo("Warnings", "\n".join(self.invalid_files))
def clear_images(self):
self.image_paths.clear()
self.invalid_files.clear()
self.refresh_listbox()
self.output_label.config(text="")
# Cleanup temporary files
for temp in self.temp_files:
try:
os.remove(temp)
except:
pass
self.temp_files.clear()
def refresh_listbox(self):
self.listbox.delete(0, tk.END)
for f in self.image_paths:
self.listbox.insert(tk.END, os.path.basename(f))
# ---------------- OCR Extraction ----------------
def start_extraction(self):
if not self.image_paths:
messagebox.showwarning("Extract Text", "No images uploaded!")
return
threading.Thread(target=self.extract_text, daemon=True).start()
def extract_text(self):
lang_name = self.lang_var.get()
lang_code = self.OCR_LANGUAGES.get(lang_name, "eng")
all_text = ""
total = len(self.image_paths)
for idx, path in enumerate(self.image_paths, 1):
try:
img = Image.open(path)
try:
text = pytesseract.image_to_string(img, lang=lang_code)
if not text.strip():
text = f"No text detected in '{os.path.basename(path)}'."
except:
text = f"No text detected in '{os.path.basename(path)}'."
all_text += f"--- {os.path.basename(path)} ---\n{text}\n\n"
except:
text = f"No text detected in '{os.path.basename(path)}'."
all_text += f"--- {os.path.basename(path)} ---\n{text}\n\n"
self.progress_var.set(idx / total * 100)
time.sleep(0.05)
self.output_label.config(text=all_text)
self.progress_var.set(0)
# ---------------- Copy & Save ----------------
def copy_text(self):
if self.output_label.cget("text").strip():
pyperclip.copy(self.output_label.cget("text"))
messagebox.showinfo("Copy", "Copied to clipboard!")
else:
messagebox.showwarning("Copy", "No text to copy!")
def save_dialog(self):
if not self.output_label.cget("text").strip():
messagebox.showwarning("Save", "Nothing to save!")
return
option = messagebox.askyesnocancel("Save Format",
"Choose format:\n\nYes = PDF\nNo = Word (.docx)\nCancel = TXT")
if option is True: # PDF
self.save_pdf()
elif option is False: # DOCX
self.save_docx()
elif option is None: # TXT
self.save_txt()
def save_txt(self):
file_path = filedialog.asksaveasfilename(defaultextension=".txt",
filetypes=[("Text Files","*.txt")])
if not file_path:
return
try:
with open(file_path, "w", encoding="utf-8") as f:
f.write(self.output_label.cget("text"))
messagebox.showinfo("Saved", "Saved as TXT!")
except Exception as e:
messagebox.showerror("Error", f"Could not save TXT:\n{e}")
def save_pdf(self):
file_path = filedialog.asksaveasfilename(defaultextension=".pdf",
filetypes=[("PDF Files","*.pdf")])
if not file_path:
return
text = self.output_label.cget("text")
doc = SimpleDocTemplate(file_path, pagesize=A4, rightMargin=40, leftMargin=40,
topMargin=40, bottomMargin=40)
styles = getSampleStyleSheet()
story = []
for line in text.split("\n"):
story.append(Paragraph(line.replace(" ", " "), styles["Normal"]))
story.append(Spacer(1, 4))
try:
doc.build(story)
messagebox.showinfo("Saved", "Saved as PDF!")
except Exception as e:
messagebox.showerror("Error", f"Could not save PDF:\n{e}")
def save_docx(self):
file_path = filedialog.asksaveasfilename(defaultextension=".docx",
filetypes=[("Word Files","*.docx")])
if not file_path:
return
try:
doc = Document()
doc.add_paragraph(self.output_label.cget("text"))
doc.save(file_path)
messagebox.showinfo("Saved", "Saved as Word (.docx)!")
except Exception as e:
messagebox.showerror("Error", f"Could not save Word:\n{e}")
# ---------------- Run ----------------
if __name__ == "__main__":
root = TkinterDnD.Tk()
SplashScreen(root)
app = ExtractMateApp(root)
root.mainloop()
📌 Nothing is omitted
📌 No refactoring
📌 No placeholders
📌 No hidden dependencies
Just run:
python extractmate.py
✨ Key Features Explained
🖱️ Drag & Drop Support
Powered by tkinterDnD2, users can drop images anywhere in the window.
🌍 Multi-Language OCR
Switch languages dynamically using a dropdown mapped to Tesseract language packs.
🖼️ SVG Compatibility
SVG files are converted into temporary PNGs using CairoSVG, then cleaned up automatically.
🚀 Threaded OCR
OCR runs in a background thread to keep the UI responsive.
💾 Export Options
Save extracted text as:
.txt
.docx
🧠 Why This Project Is Useful
This project touches real-world Python topics:
GUI architecture
Thread safety
File validation & limits
UX feedback (progress bars)
Temporary file handling
Cross-format exporting
It’s perfect if you’re learning:
Desktop app development
OCR workflows
Tkinter beyond basics
🔮 Ideas for Improvements
📄 PDF input OCR
📷 Webcam OCR
🧠 Layout-aware OCR
🌍 Auto language detection
📦 PyInstaller builds for Windows/macOS
🧡 Final Thoughts
ExtractMate proves you can build modern-looking, practical desktop tools with Python.
Top comments (0)