diff --git a/tripit_add.py b/tripit_add.py
new file mode 100644
index 0000000..f8a87ae
--- /dev/null
+++ b/tripit_add.py
@@ -0,0 +1,142 @@
+#!/usr/bin/python3
+
+from collections import defaultdict
+from datetime import datetime
+from typing import List, Tuple
+
+import lxml.etree as etree
+import lxml.builder as builder
+import more_itertools as mlt
+from pytz import timezone
+from requests_oauthlib import OAuth1Session
+import simplejson as json
+
+E = builder.ElementMaker()
+
+Request = E.Request
+Segment = E.Segment
+StartDateTime = E.StartDateTime
+EndDateTime = E.EndDateTime
+RailObject = E.RailObject
+date_ = E.date
+time_ = E.time
+
+
+def get_tokens(filename: str) -> Tuple[str]:
+
+    with open(filename) as handle:
+        tokens = json.load(handle)
+
+    return (tokens["client_key"], tokens["client_secret"],
+            tokens["resource_owner_key"], tokens["resource_owner_secret"])
+
+
+def add_train_segment(start_station: str,
+                      end_station: str,
+                      start_time: str,
+                      end_time: str,
+                      trip_date: str,
+                      train: str = None,
+                      number: int = None,
+                      carrier: str = None) -> Segment:
+
+    start_station_name = E.start_station_name
+    end_station_name = E.end_station_name
+    train_type = E.train_type
+    train_number = E.train_number
+    carrier_name = E.carrier_name
+
+    data = []
+
+    data.append(StartDateTime(date_(trip_date), time_(start_time)))
+    data.append(EndDateTime(date_(trip_date), time_(end_time)))
+    data.append(start_station_name(start_station))
+    data.append(end_station_name(end_station))
+
+    if number:
+        data.append(train_number(number))
+
+    if carrier:
+        data.append(carrier_name(carrier))
+
+    if train_type:
+        data.append(train_type(train))
+
+    doc = Segment(*data)
+
+    return doc
+
+
+def _convert_datetimes(segment: int) -> List[str]:
+
+    date_time = datetime.fromtimestamp(segment,
+                                       timezone("Japan"))
+
+    date = date_time.strftime("%Y-%m-%d")
+    time = date_time.strftime("%H:%M:00")
+
+    return date, time
+
+
+def parse_json(json_data: dict, carrier: str = None,
+               result_number: int = None) -> str:
+
+    jdoc = json_data
+
+    result = jdoc["result"]
+    record = result[result_number - 1]
+    segments = list()
+
+    iter_records = mlt.peekable(record["steps"])
+
+    previous_record = defaultdict(None)
+
+    for segment in iter_records:
+
+        start_station = segment["start_station"].title()
+        start_date, start_time = _convert_datetimes(segment["start_time"])
+
+        end_station = segment["end_station"].title()
+        end_date, end_time = _convert_datetimes(segment["end_time"])
+
+        # FIXME: Impossible to extract it from the current Hyperdia data
+        carrier = "" if carrier is None else carrier
+        train_type = segment["train_name"]
+        train_number = ("" if segment.get("train_number") is None
+                        else segment["train_number"])
+
+        next_record = iter_records.peek(None)
+
+        if next_record is not None:
+
+            if not segment["go_through"] and next_record["go_through"]:
+                # Next one is going through: set the start,
+                # but don't add anything
+                previous_record["start_station"] = start_station
+                previous_record["start_time"] = start_time
+                previous_record["start_date"] = start_date
+                previous_record["train_name"] = train_type
+                continue
+            elif not next_record["go_through"] and segment["go_through"]:
+                # The next one is a "real" one, set the end, and add
+                # the record
+                start_station = previous_record["start_station"]
+                start_time = previous_record["start_time"]
+                start_date = previous_record["start_date"]
+                train_type = previous_record["train_name"]
+                previous_record = dict.fromkeys(previous_record.keys(), None)
+
+        subdoc = add_train_segment(start_station,
+                                   end_station,
+                                   start_time,
+                                   end_time,
+                                   start_date,
+                                   train=train_type,
+                                   number=train_number,
+                                   carrier=carrier)
+        segments.append(subdoc)
+
+    rail_object = RailObject(*segments)
+    request_object = Request(rail_object)
+
+    return etree.tounicode(request_object)