diff --git a/src/file_manipulator.py b/src/file_manipulator.py
index b7815cc..fe38f21 100644
--- a/src/file_manipulator.py
+++ b/src/file_manipulator.py
@@ -1,8 +1,11 @@
 import os
 from src.filler import Filler
 from src.llm import LLM
+from src.logger import setup_logger
 from commonforms import prepare_form
 
+logger = setup_logger(__name__)
+
 
 class FileManipulator:
     def __init__(self):
@@ -22,26 +25,21 @@ def fill_form(self, user_input: str, fields: list, pdf_form_path: str):
         It receives the raw data, runs the PDF filling logic,
         and returns the path to the newly created file.
         """
-        print("[1] Received request from frontend.")
-        print(f"[2] PDF template path: {pdf_form_path}")
+        logger.info("Received request from frontend.")
+        logger.info(f"PDF template path: {pdf_form_path}")
 
         if not os.path.exists(pdf_form_path):
-            print(f"Error: PDF template not found at {pdf_form_path}")
-            return None  # Or raise an exception
+            logger.error(f"PDF template not found at {pdf_form_path}")
+            return None
 
-        print("[3] Starting extraction and PDF filling process...")
+        logger.info("Starting extraction and PDF filling process...")
         try:
             self.llm._target_fields = fields
             self.llm._transcript_text = user_input
             output_name = self.filler.fill_form(pdf_form=pdf_form_path, llm=self.llm)
-
-            print("\n----------------------------------")
-            print("✅ Process Complete.")
-            print(f"Output saved to: {output_name}")
-
+            logger.info("Process Complete.")
+            logger.info(f"Output saved to: {output_name}")
             return output_name
-
         except Exception as e:
-            print(f"An error occurred during PDF generation: {e}")
-            # Re-raise the exception so the frontend can handle it
-            raise e
+            logger.error(f"An error occurred during PDF generation: {e}")
+            raise e
\ No newline at end of file
diff --git a/src/llm.py b/src/llm.py
index 70937f9..ab370e1 100644
--- a/src/llm.py
+++ b/src/llm.py
@@ -1,15 +1,19 @@
 import json
 import os
+import re
 import requests
 
+from src.logger import setup_logger
+logger = setup_logger(__name__)
+
 
 class LLM:
     def __init__(self, transcript_text=None, target_fields=None, json=None):
         if json is None:
             json = {}
-        self._transcript_text = transcript_text  # str
-        self._target_fields = target_fields  # List, contains the template field.
-        self._json = json  # dictionary
+        self._transcript_text = transcript_text
+        self._target_fields = target_fields
+        self._json = json
 
     def type_check_all(self):
         if type(self._transcript_text) is not str:
@@ -24,39 +28,31 @@ def type_check_all(self):
             )
 
     def build_prompt(self, current_field):
-        """
-        This method is in charge of the prompt engineering. It creates a specific prompt for each target field.
-        @params: current_field -> represents the current element of the json that is being prompted.
-        """
-        prompt = f""" 
+        prompt = f"""
             SYSTEM PROMPT:
-            You are an AI assistant designed to help fillout json files with information extracted from transcribed voice recordings. 
-            You will receive the transcription, and the name of the JSON field whose value you have to identify in the context. Return 
-            only a single string containing the identified value for the JSON field. 
+            You are an AI assistant designed to help fillout json files with information extracted from transcribed voice recordings.
+            You will receive the transcription, and the name of the JSON field whose value you have to identify in the context. Return
+            only a single string containing the identified value for the JSON field.
             If the field name is plural, and you identify more than one possible value in the text, return both separated by a ";".
             If you don't identify the value in the provided text, return "-1".
             ---
             DATA:
             Target JSON field to find in text: {current_field}
-            
+
             TEXT: {self._transcript_text}
             """
-
         return prompt
 
     def main_loop(self):
-        # self.type_check_all()
         for field in self._target_fields.keys():
             prompt = self.build_prompt(field)
-            # print(prompt)
-            # ollama_url = "http://localhost:11434/api/generate"
             ollama_host = os.getenv("OLLAMA_HOST", "http://localhost:11434").rstrip("/")
             ollama_url = f"{ollama_host}/api/generate"
 
             payload = {
                 "model": "mistral",
                 "prompt": prompt,
-                "stream": False,  # don't really know why --> look into this later.
+                "stream": False,
             }
 
             try:
@@ -70,23 +66,23 @@ def main_loop(self):
             except requests.exceptions.HTTPError as e:
                 raise RuntimeError(f"Ollama returned an error: {e}")
 
-            # parse response
             json_data = response.json()
             parsed_response = json_data["response"]
-            # print(parsed_response)
             self.add_response_to_json(field, parsed_response)
 
-        print("----------------------------------")
-        print("\t[LOG] Resulting JSON created from the input text:")
-        print(json.dumps(self._json, indent=2))
-        print("--------- extracted data ---------")
+        # ── NEW: build confidence-scored result ──
+        extraction_result = self.build_extraction_result()
+        logger.info("----------------------------------")
+        logger.info("Resulting JSON with confidence scores:")
+        logger.info(json.dumps(extraction_result, indent=2))
+        logger.info("--------- extracted data ---------")
 
         return self
 
     def add_response_to_json(self, field, value):
         """
-        this method adds the following value under the specified field,
-        or under a new field if the field doesn't exist, to the json dict
+        Adds the value under the specified field.
+        Existing PR #337 fix preserved — handles non-list fields safely.
         """
         value = value.strip().replace('"', "")
         parsed_value = None
@@ -97,39 +93,98 @@ def add_response_to_json(self, field, value):
         if ";" in value:
             parsed_value = self.handle_plural_values(value)
 
-        if field in self._json.keys():
-            self._json[field].append(parsed_value)
+        if field in self._json:
+            if isinstance(self._json[field], list):
+                self._json[field].append(parsed_value)
+            else:
+                self._json[field] = [self._json[field], parsed_value]
         else:
             self._json[field] = parsed_value
 
         return
 
     def handle_plural_values(self, plural_value):
-        """
-        This method handles plural values.
-        Takes in strings of the form 'value1; value2; value3; ...; valueN'
-        returns a list with the respective values -> [value1, value2, value3, ..., valueN]
-        """
         if ";" not in plural_value:
             raise ValueError(
                 f"Value is not plural, doesn't have ; separator, Value: {plural_value}"
             )
 
-        print(
-            f"\t[LOG]: Formating plural values for JSON, [For input {plural_value}]..."
-        )
+        logger.info(f"Formatting plural values for JSON, input: {plural_value}")
         values = plural_value.split(";")
 
-        # Remove trailing leading whitespace
         for i in range(len(values)):
             current = i + 1
             if current < len(values):
                 clean_value = values[current].lstrip()
                 values[current] = clean_value
 
-        print(f"\t[LOG]: Resulting formatted list of values: {values}")
-
+        logger.info(f"Resulting formatted list of values: {values}")
         return values
 
+    def _compute_field_confidence(self, value) -> float:
+        """
+        Heuristic confidence scoring for an extracted field value.
+        Returns a float between 0.0 and 1.0.
+        """
+        if value is None or value == "" or value == "-1":
+            return 0.0
+        if isinstance(value, list):
+            return 0.85 if len(value) > 0 else 0.0
+        if isinstance(value, str):
+            vague_patterns = [
+                r"not (specified|mentioned|provided|found|available)",
+                r"^n/?a$",
+                r"^\?+$",
+                r"^unknown$"
+            ]
+            for pattern in vague_patterns:
+                if re.search(pattern, value.strip(), re.IGNORECASE):
+                    return 0.2
+            if len(value.strip()) < 3:
+                return 0.3
+            return 0.9
+        return 0.8
+
+    def build_extraction_result(self) -> dict:
+        """
+        Wraps each extracted field with a confidence score.
+        Adds top-level _meta block with requires_review flag
+        when any field confidence is below threshold.
+        """
+        CONFIDENCE_THRESHOLD = 0.5
+        result = {}
+        low_confidence_fields = []
+
+        for field, value in self._json.items():
+            score = self._compute_field_confidence(value)
+            result[field] = {
+                "value": value,
+                "confidence": round(score, 2)
+            }
+            if score < CONFIDENCE_THRESHOLD:
+                low_confidence_fields.append(field)
+
+        total_fields = len(self._json)
+        overall = (
+            round(sum(result[f]["confidence"] for f in result) / total_fields, 2)
+            if total_fields > 0 else 0.0
+        )
+
+        result["_meta"] = {
+            "requires_review": len(low_confidence_fields) > 0,
+            "low_confidence_fields": low_confidence_fields,
+            "overall_confidence": overall
+        }
+
+        if result["_meta"]["requires_review"]:
+            logger.warning(
+                "Extraction requires human review. Low confidence fields: %s",
+                low_confidence_fields
+            )
+        else:
+            logger.info("Extraction complete. All fields passed confidence threshold.")
+
+        return result
+
     def get_data(self):
-        return self._json
+        return self._json
\ No newline at end of file
diff --git a/src/logger.py b/src/logger.py
new file mode 100644
index 0000000..d8eab83
--- /dev/null
+++ b/src/logger.py
@@ -0,0 +1,25 @@
+import logging
+
+
+def setup_logger(name: str):
+    """
+    Sets up and returns a logger with the given name.
+    Avoids adding duplicate handlers if logger already exists.
+    """
+    logger = logging.getLogger(name)
+
+    if logger.handlers:
+        return logger
+
+    logger.setLevel(logging.INFO)
+
+    handler = logging.StreamHandler()
+
+    formatter = logging.Formatter(
+        "%(asctime)s | %(levelname)s | %(name)s | %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S"
+    )
+    handler.setFormatter(formatter)
+    logger.addHandler(handler)
+
+    return logger
\ No newline at end of file
diff --git a/src/test/test_confidence.py b/src/test/test_confidence.py
new file mode 100644
index 0000000..908d76f
--- /dev/null
+++ b/src/test/test_confidence.py
@@ -0,0 +1,77 @@
+import sys
+import os
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '../../')))
+
+import pytest
+from src.llm import LLM
+
+
+@pytest.fixture
+def llm():
+    fields = {"name": None, "date": None, "location": None}
+    return LLM(transcript_text="John went to Mumbai on 2024-01-15", target_fields=fields)
+
+
+# --- _compute_field_confidence tests ---
+
+def test_confidence_none_value(llm):
+    assert llm._compute_field_confidence(None) == 0.0
+
+def test_confidence_empty_string(llm):
+    assert llm._compute_field_confidence("") == 0.0
+
+def test_confidence_minus_one(llm):
+    assert llm._compute_field_confidence("-1") == 0.0
+
+def test_confidence_normal_string(llm):
+    assert llm._compute_field_confidence("John Smith") >= 0.8
+
+def test_confidence_vague_not_specified(llm):
+    assert llm._compute_field_confidence("not specified") < 0.5
+
+def test_confidence_vague_na(llm):
+    assert llm._compute_field_confidence("N/A") < 0.5
+
+def test_confidence_vague_unknown(llm):
+    assert llm._compute_field_confidence("unknown") < 0.5
+
+def test_confidence_short_string(llm):
+    assert llm._compute_field_confidence("ab") < 0.5
+
+def test_confidence_plural_list(llm):
+    assert llm._compute_field_confidence(["val1", "val2"]) == 0.85
+
+def test_confidence_empty_list(llm):
+    assert llm._compute_field_confidence([]) == 0.0
+
+
+# --- build_extraction_result tests ---
+
+def test_requires_review_true_when_low_confidence(llm):
+    llm._json = {"name": "John", "date": None, "location": "N/A"}
+    result = llm.build_extraction_result()
+    assert result["_meta"]["requires_review"] is True
+    assert "date" in result["_meta"]["low_confidence_fields"]
+
+def test_requires_review_false_when_all_confident(llm):
+    llm._json = {"name": "John Smith", "date": "2024-01-15", "location": "Mumbai"}
+    result = llm.build_extraction_result()
+    assert result["_meta"]["requires_review"] is False
+    assert result["_meta"]["low_confidence_fields"] == []
+
+def test_overall_confidence_is_average(llm):
+    llm._json = {"name": "Alice", "date": "2024-03-01"}
+    result = llm.build_extraction_result()
+    assert 0.0 <= result["_meta"]["overall_confidence"] <= 1.0
+
+def test_each_field_has_value_and_confidence(llm):
+    llm._json = {"name": "Bob"}
+    result = llm.build_extraction_result()
+    assert "value" in result["name"]
+    assert "confidence" in result["name"]
+
+def test_empty_json_returns_zero_confidence(llm):
+    llm._json = {}
+    result = llm.build_extraction_result()
+    assert result["_meta"]["overall_confidence"] == 0.0
+    assert result["_meta"]["requires_review"] is False
\ No newline at end of file