diff --git a/pyproject.toml b/pyproject.toml
index 887ada47..e94a7da4 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -42,7 +42,6 @@ dependencies = [
     "scikit-learn >= 1.5",
     "xgboost-cpu >= 2.1.4", # regular xgboost includes nvidia libraries which bloat the package size on linux. For PyProphet, we likely would not need GPU support.
     "matplotlib",
-    "pyarrow",
     "pypdf",
     "psutil",
     "pyopenms",
@@ -55,6 +54,7 @@ dependencies = [
 testing = ["pytest", "pytest-regtest", "pytest-xdist"]
 docs = ["sphinx", "sphinx-copybutton", "sphinx_rtd_theme", "pydata_sphinx_theme", "sphinx-click"]
 dev = ["pyprophet[testing]", "pyprophet[docs]", "black", "ruff", "mypy"]
+parquet = ["pyarrow"]
 
 # Define console entry points
 [project.scripts]
diff --git a/pyprophet/io/__init__.py b/pyprophet/io/__init__.py
index 6bbbfa4b..372e3726 100644
--- a/pyprophet/io/__init__.py
+++ b/pyprophet/io/__init__.py
@@ -15,7 +15,7 @@
 Dependencies:
 -------------
 - `pandas`
-- `pyarrow`
+- `pyarrow`(optional, for Parquet support)
 - `duckdb`
 - `sqlite3`
 - `loguru`
diff --git a/pyprophet/io/dispatcher.py b/pyprophet/io/dispatcher.py
index 00c4eb56..4aeeeaff 100644
--- a/pyprophet/io/dispatcher.py
+++ b/pyprophet/io/dispatcher.py
@@ -16,52 +16,28 @@
 
 from loguru import logger
 
+from .util import (
+    _get_parquet_reader_class_for_config,
+    _get_parquet_writer_class_for_config,
+)
 from .._config import ExportIOConfig, IPFIOConfig, LevelContextIOConfig, RunnerIOConfig
 
 # Export I/O
 from .export.osw import OSWReader as ExportOSWReader
 from .export.osw import OSWWriter as ExportOSWWriter
 from .export.sqmass import SqMassWriter as ExportSqMassWriter
-from .export.parquet import (
-    ParquetReader as ExportParquetReader,
-)
-from .export.parquet import (
-    ParquetWriter as ExportParquetWriter,
-)
-from .export.split_parquet import (
-    SplitParquetReader as ExportSplitParquetReader,
-)
-from .export.split_parquet import (
-    SplitParquetWriter as ExportSplitParquetWriter,
-)
 
 # IPF I/O
 from .ipf.osw import OSWReader as IPFOSWReader
 from .ipf.osw import OSWWriter as IPFOSWWriter
-from .ipf.parquet import ParquetReader as IPFParquetReader
-from .ipf.parquet import ParquetWriter as IPFParquetWriter
-from .ipf.split_parquet import SplitParquetReader as IPFSplitParquetReader
-from .ipf.split_parquet import SplitParquetWriter as IPFSplitParquetWriter
 
 # Levels Context I/O
 from .levels_context.osw import OSWReader as LevelContextOSWReader
 from .levels_context.osw import OSWWriter as LevelContextOSWWriter
-from .levels_context.parquet import ParquetReader as LevelContextParquetReader
-from .levels_context.parquet import ParquetWriter as LevelContextParquetWriter
-from .levels_context.split_parquet import (
-    SplitParquetReader as LevelContextSplitParquetReader,
-)
-from .levels_context.split_parquet import (
-    SplitParquetWriter as LevelContextSplitParquetWriter,
-)
 
 # Scoring I/O
 from .scoring.osw import OSWReader as ScoringOSWReader
 from .scoring.osw import OSWWriter as ScoringOSWWriter
-from .scoring.parquet import ParquetReader as ParquetScoringReader
-from .scoring.parquet import ParquetWriter as ParquetScoringWriter
-from .scoring.split_parquet import SplitParquetReader as SplitParquetScoringReader
-from .scoring.split_parquet import SplitParquetWriter as SplitParquetScoringWriter
 from .scoring.tsv import TSVReader as ScoringTSVReader
 from .scoring.tsv import TSVWriter as ScoringTSVWriter
 
@@ -123,29 +99,13 @@ def _get_osw_reader(config):
 
     @staticmethod
     def _get_parquet_reader(config):
-        if isinstance(config, RunnerIOConfig):
-            return ParquetScoringReader(config)
-        elif isinstance(config, IPFIOConfig):
-            return IPFParquetReader(config)
-        elif isinstance(config, LevelContextIOConfig):
-            return LevelContextParquetReader(config)
-        elif isinstance(config, ExportIOConfig):
-            return ExportParquetReader(config)
-        else:
-            raise ValueError(f"Unsupported config context: {type(config).__name__}")
+        cls = _get_parquet_reader_class_for_config(config, split=False)
+        return cls(config)
 
     @staticmethod
     def _get_split_parquet_reader(config):
-        if isinstance(config, RunnerIOConfig):
-            return SplitParquetScoringReader(config)
-        elif isinstance(config, IPFIOConfig):
-            return IPFSplitParquetReader(config)
-        elif isinstance(config, LevelContextIOConfig):
-            return LevelContextSplitParquetReader(config)
-        elif isinstance(config, ExportIOConfig):
-            return ExportSplitParquetReader(config)
-        else:
-            raise ValueError(f"Unsupported config context: {type(config).__name__}")
+        cls = _get_parquet_reader_class_for_config(config, split=True)
+        return cls(config)
 
     @staticmethod
     def _get_tsv_reader(config):
@@ -223,29 +183,13 @@ def _get_sqmass_writer(config):
 
     @staticmethod
     def _get_parquet_writer(config):
-        if isinstance(config, RunnerIOConfig):
-            return ParquetScoringWriter(config)
-        elif isinstance(config, IPFIOConfig):
-            return IPFParquetWriter(config)
-        elif isinstance(config, LevelContextIOConfig):
-            return LevelContextParquetWriter(config)
-        elif isinstance(config, ExportIOConfig):
-            return ExportParquetWriter(config)
-        else:
-            raise ValueError(f"Unsupported config context: {type(config).__name__}")
+        cls = _get_parquet_writer_class_for_config(config, split=False)
+        return cls(config)
 
     @staticmethod
     def _get_split_parquet_writer(config):
-        if isinstance(config, RunnerIOConfig):
-            return SplitParquetScoringWriter(config)
-        elif isinstance(config, IPFIOConfig):
-            return IPFSplitParquetWriter(config)
-        elif isinstance(config, LevelContextIOConfig):
-            return LevelContextSplitParquetWriter(config)
-        elif isinstance(config, ExportIOConfig):
-            return ExportSplitParquetWriter(config)
-        else:
-            raise ValueError(f"Unsupported config context: {type(config).__name__}")
+        cls = _get_parquet_writer_class_for_config(config, split=True)
+        return cls(config)
 
     @staticmethod
     def _get_tsv_writer(config):
diff --git a/pyprophet/io/ipf/parquet.py b/pyprophet/io/ipf/parquet.py
index f9c7ff93..f97e6bde 100644
--- a/pyprophet/io/ipf/parquet.py
+++ b/pyprophet/io/ipf/parquet.py
@@ -1,14 +1,17 @@
 import os
-from typing import Literal
 from shutil import copyfile
-import pandas as pd
-import pyarrow as pa
-import duckdb
+from typing import Literal
+
 import click
+import duckdb
+import pandas as pd
 from loguru import logger
-from ..util import get_parquet_column_names
-from .._base import BaseParquetReader, BaseParquetWriter
+
 from ..._config import IPFIOConfig
+from .._base import BaseParquetReader, BaseParquetWriter
+from ..util import _ensure_pyarrow, get_parquet_column_names
+
+pa, _, _ = _ensure_pyarrow()
 
 
 class ParquetReader(BaseParquetReader):
diff --git a/pyprophet/io/ipf/split_parquet.py b/pyprophet/io/ipf/split_parquet.py
index acb733c3..085cdea0 100644
--- a/pyprophet/io/ipf/split_parquet.py
+++ b/pyprophet/io/ipf/split_parquet.py
@@ -1,16 +1,17 @@
-import os
 import glob
-from shutil import copyfile
+import os
 from typing import Literal
-import pandas as pd
-import pyarrow as pa
-import duckdb
+
 import click
+import duckdb
+import pandas as pd
 from loguru import logger
 
-from ..util import get_parquet_column_names
-from .._base import BaseSplitParquetReader, BaseSplitParquetWriter
 from ..._config import IPFIOConfig
+from .._base import BaseSplitParquetReader, BaseSplitParquetWriter
+from ..util import _ensure_pyarrow, get_parquet_column_names
+
+pa, _, _ = _ensure_pyarrow()
 
 
 class SplitParquetReader(BaseSplitParquetReader):
diff --git a/pyprophet/io/levels_context/parquet.py b/pyprophet/io/levels_context/parquet.py
index 49e33f3c..24f2c3e8 100644
--- a/pyprophet/io/levels_context/parquet.py
+++ b/pyprophet/io/levels_context/parquet.py
@@ -1,14 +1,15 @@
-import os
-from typing import Literal
 from shutil import copyfile
-import pandas as pd
-import pyarrow as pa
-import duckdb
+
 import click
+import duckdb
+import pandas as pd
 from loguru import logger
-from ..util import get_parquet_column_names
-from .._base import BaseParquetReader, BaseParquetWriter
+
 from ..._config import LevelContextIOConfig
+from .._base import BaseParquetReader, BaseParquetWriter
+from ..util import _ensure_pyarrow, get_parquet_column_names
+
+pa, _, _ = _ensure_pyarrow()
 
 
 class ParquetReader(BaseParquetReader):
diff --git a/pyprophet/io/levels_context/split_parquet.py b/pyprophet/io/levels_context/split_parquet.py
index 671ff88f..3f0812d3 100644
--- a/pyprophet/io/levels_context/split_parquet.py
+++ b/pyprophet/io/levels_context/split_parquet.py
@@ -1,18 +1,16 @@
-import os
 import glob
-from shutil import copyfile
-from typing import Literal
-import pandas as pd
-import pyarrow as pa
-import duckdb
+import os
+
 import click
+import duckdb
+import pandas as pd
 from loguru import logger
 
-from .._base import BaseSplitParquetReader, BaseSplitParquetWriter
 from ..._config import LevelContextIOConfig
-from ..util import (
-    get_parquet_column_names,
-)
+from .._base import BaseSplitParquetReader, BaseSplitParquetWriter
+from ..util import _ensure_pyarrow, get_parquet_column_names
+
+pa, _, _ = _ensure_pyarrow()
 
 
 class SplitParquetReader(BaseSplitParquetReader):
diff --git a/pyprophet/io/scoring/parquet.py b/pyprophet/io/scoring/parquet.py
index 538832fa..b1c3626c 100644
--- a/pyprophet/io/scoring/parquet.py
+++ b/pyprophet/io/scoring/parquet.py
@@ -1,14 +1,17 @@
 import sys
 from shutil import copyfile
+
+import click
+import duckdb
 import pandas as pd
 import polars as pl
-import pyarrow as pa
-import duckdb
-import click
 from loguru import logger
-from ..util import get_parquet_column_names
-from .._base import BaseParquetReader, BaseParquetWriter, RowCountMismatchError
+
 from ..._config import RunnerIOConfig
+from .._base import BaseParquetReader, BaseParquetWriter, RowCountMismatchError
+from ..util import _ensure_pyarrow, get_parquet_column_names
+
+pa, _, _ = _ensure_pyarrow()
 
 
 class ParquetReader(BaseParquetReader):
diff --git a/pyprophet/io/scoring/split_parquet.py b/pyprophet/io/scoring/split_parquet.py
index 2e5eb464..de09e720 100644
--- a/pyprophet/io/scoring/split_parquet.py
+++ b/pyprophet/io/scoring/split_parquet.py
@@ -1,16 +1,15 @@
 import os
-import sys
-import glob
-from shutil import copyfile
-import pandas as pd
-import pyarrow as pa
-import duckdb
+
 import click
+import duckdb
+import pandas as pd
 from loguru import logger
 
-from ..util import get_parquet_column_names
-from .._base import BaseSplitParquetReader, BaseSplitParquetWriter
 from ..._config import RunnerIOConfig
+from .._base import BaseSplitParquetReader, BaseSplitParquetWriter
+from ..util import _ensure_pyarrow, get_parquet_column_names
+
+pa, _, _ = _ensure_pyarrow()
 
 
 class SplitParquetReader(BaseSplitParquetReader):
diff --git a/pyprophet/io/util.py b/pyprophet/io/util.py
index 4d07617d..9cce1d95 100644
--- a/pyprophet/io/util.py
+++ b/pyprophet/io/util.py
@@ -44,13 +44,40 @@
 import os
 from collections import defaultdict
 import sqlite3
+import importlib
+from typing import Type
 import duckdb
 import click
 import pandas as pd
-import pyarrow.parquet as pq
 import pyopenms as poms
 from loguru import logger
-from pyarrow.lib import ArrowInvalid, ArrowIOError
+
+
+def _ensure_pyarrow():
+    """
+    Avoid importing pyarrow at module import time; import lazily in functions that need it.
+    """
+    try:
+        import pyarrow as pa  # pylint: disable=C0415
+
+        # Ensure the parquet submodule is loaded and available as pa.parquet
+        try:
+            import pyarrow.parquet as pq  # pylint: disable=C0415
+        except Exception:
+            pq = None
+        from pyarrow.lib import ArrowInvalid, ArrowIOError  # pylint: disable=C0415
+
+        # Some pyarrow installs don't automatically expose the parquet submodule
+        # as an attribute on the top-level module until it's imported. Attach it
+        # so callers can use pa.parquet.*
+        if pq is not None and not hasattr(pa, "parquet"):
+            setattr(pa, "parquet", pq)
+
+        return pa, ArrowInvalid, ArrowIOError
+    except ImportError as exc:
+        raise click.ClickException(
+            "Parquet support requires 'pyarrow'. Install with 'pip install pyarrow' or 'pip install pyprophet[parquet]'."
+        ) from exc
 
 
 def is_tsv_file(file_path):
@@ -248,6 +275,60 @@ def create_index_if_not_exists(con, index_name, table_name, column_name):
         con.execute(f"CREATE INDEX {index_name} ON {table_name} ({column_name})")
 
 
+def _lazy_parquet_class(module_path: str, class_name: str) -> Type:
+    """
+    Import the given module (relative to this package) and return the class.
+    Raises a ClickException with a friendly message if the import fails (e.g. missing pyarrow).
+    """
+    try:
+        mod = importlib.import_module(module_path, package=__package__)
+        return getattr(mod, class_name)
+    except ModuleNotFoundError as exc:
+        # Likely pyarrow or the module itself is missing; user should install the parquet extra.
+        raise click.ClickException(
+            "Parquet support requires the 'pyarrow' package. "
+            "Install it with 'pip install pyarrow' or 'pip install pyprophet[parquet]'."
+        ) from exc
+    except Exception:
+        # Propagate other exceptions (syntax errors, attribute errors) to surface the real problem.
+        raise
+
+
+def _area_from_config(config) -> str:
+    """
+    Map a config instance to its package area name used in the io package.
+    """
+    # Avoid importing config classes here to prevent circular imports.
+    cname = type(config).__name__
+    if cname == "RunnerIOConfig":
+        return "scoring"
+    if cname == "IPFIOConfig":
+        return "ipf"
+    if cname == "LevelContextIOConfig":
+        return "levels_context"
+    if cname == "ExportIOConfig":
+        return "export"
+    raise ValueError(f"Unsupported config context: {type(config).__name__}")
+
+
+def _get_parquet_reader_class_for_config(config, split: bool = False) -> Type:
+    _, _, _ = _ensure_pyarrow()
+    area = _area_from_config(config)
+    module = f".{area}.split_parquet" if split else f".{area}.parquet"
+    return _lazy_parquet_class(
+        module, "SplitParquetReader" if split else "ParquetReader"
+    )
+
+
+def _get_parquet_writer_class_for_config(config, split: bool = False) -> Type:
+    _, _, _ = _ensure_pyarrow()
+    area = _area_from_config(config)
+    module = f".{area}.split_parquet" if split else f".{area}.parquet"
+    return _lazy_parquet_class(
+        module, "SplitParquetWriter" if split else "ParquetWriter"
+    )
+
+
 def is_parquet_file(file_path):
     """
     Check if the file is a valid Parquet file.
@@ -259,7 +340,8 @@ def is_parquet_file(file_path):
 
     # Then verify it's actually a parquet file
     try:
-        pq.read_schema(file_path)
+        pa, ArrowInvalid, ArrowIOError = _ensure_pyarrow()  # pylint: disable=C0103
+        pa.parquet.read_schema(file_path)
         return True
     except (ArrowInvalid, ArrowIOError, OSError):
         return False
@@ -327,7 +409,8 @@ def get_parquet_column_names(file_path):
     Retrieves column names from a Parquet file without reading the entire file.
     """
     try:
-        table_schema = pq.read_schema(file_path)
+        pa, _, _ = _ensure_pyarrow()
+        table_schema = pa.parquet.read_schema(file_path)
         return table_schema.names
     except Exception as e:
         print(f"An error occurred while reading schema from '{file_path}': {e}")
diff --git a/scripts/build/build_linux.sh b/scripts/build/build_linux.sh
index 29fc5e5c..06d99551 100644
--- a/scripts/build/build_linux.sh
+++ b/scripts/build/build_linux.sh
@@ -104,6 +104,21 @@ $PYTHON -m PyInstaller \
   --noconfirm \
   --log-level INFO \
   --additional-hooks-dir packaging/pyinstaller/hooks \
+  --exclude-module pyarrow \
+  --exclude-module sphinx \
+  --exclude-module sphinx_rtd_theme \
+  --exclude-module pydata_sphinx_theme \
+  --exclude-module sphinx_copybutton \
+  --exclude-module sphinx.ext \
+  --exclude-module alabaster \
+  --exclude-module babel \
+  --exclude-module docutils \
+  --exclude-module mypy \
+  --exclude-module pytest \
+  --exclude-module pytest-regtest \
+  --exclude-module pytest-xdist \
+  --exclude-module black \
+  --exclude-module ruff \
   --hidden-import=pyprophet \
   --hidden-import=pyprophet.main \
   --collect-submodules pyprophet \
diff --git a/scripts/build/build_macos.sh b/scripts/build/build_macos.sh
index 2578abea..96a74643 100755
--- a/scripts/build/build_macos.sh
+++ b/scripts/build/build_macos.sh
@@ -21,6 +21,40 @@ if [ "$(printf '%s\n' "$REQUIRED_VERSION" "$PYTHON_VERSION" | sort -V | head -n1
     exit 1
 fi
 
+# Get version: prefer GITHUB_REF_NAME (CI), then pyproject.toml, then git tag, then default
+if [ -n "${GITHUB_REF_NAME:-}" ]; then
+    VERSION="${GITHUB_REF_NAME#v}"
+else
+    # Use tomllib (Python 3.11) to read pyproject.toml
+    VERSION=$(
+        python3 - <<'PY'
+import tomllib, sys, subprocess
+try:
+    with open("pyproject.toml", "rb") as f:
+        cfg = tomllib.load(f)
+        v = cfg.get("project", {}).get("version")
+        if v:
+            print(v)
+            sys.exit(0)
+except Exception:
+    pass
+# fallback to git tag
+try:
+    tag = subprocess.check_output(["git", "describe", "--tags", "--abbrev=0"], stderr=subprocess.DEVNULL).decode().strip()
+    print(tag.lstrip("v"))
+    sys.exit(0)
+except Exception:
+    print("0.0.0")
+PY
+    )
+fi
+
+# sanitize for filenames
+VERSION_SAFE="${VERSION//\//-}"
+VERSION_SAFE="${VERSION_SAFE// /-}"
+VERSION_SAFE="$(echo "$VERSION_SAFE" | tr -cd 'A-Za-z0-9._-')"
+
+
 # Install/upgrade build dependencies
 python3 -m pip install --upgrade pip setuptools wheel cython numpy pyinstaller
 
@@ -78,6 +112,7 @@ python3 -m PyInstaller \
   --strip \
   --log-level INFO \
   --additional-hooks-dir packaging/pyinstaller/hooks \
+  --exclude-module pyarrow \
   --exclude-module sphinx \
   --exclude-module sphinx_rtd_theme \
   --exclude-module pydata_sphinx_theme \
diff --git a/scripts/build/build_windows.bat b/scripts/build/build_windows.bat
index 0c7f9b1c..ba9c205c 100644
--- a/scripts/build/build_windows.bat
+++ b/scripts/build/build_windows.bat
@@ -61,6 +61,21 @@ python -m PyInstaller ^
     --name pyprophet ^
     --log-level INFO ^
     --additional-hooks-dir packaging/pyinstaller/hooks ^
+    --exclude-module pyarrow ^
+    --exclude-module sphinx ^
+    --exclude-module sphinx_rtd_theme ^
+    --exclude-module pydata_sphinx_theme ^
+    --exclude-module sphinx_copybutton ^
+    --exclude-module sphinx.ext ^
+    --exclude-module alabaster ^
+    --exclude-module babel ^
+    --exclude-module docutils ^
+    --exclude-module mypy ^
+    --exclude-module pytest ^
+    --exclude-module pytest-regtest ^
+    --exclude-module pytest-xdist ^
+    --exclude-module black ^
+    --exclude-module ruff ^
     --hidden-import=pyprophet ^
     --hidden-import=pyprophet.main ^
     --collect-submodules pyprophet ^
diff --git a/scripts/build/create_macos_dmg.sh b/scripts/build/create_macos_dmg.sh
index b1b8a45b..e97cb625 100644
--- a/scripts/build/create_macos_dmg.sh
+++ b/scripts/build/create_macos_dmg.sh
@@ -8,15 +8,38 @@ echo "============================================"
 echo "Creating macOS DMG Installer"
 echo "============================================"
 
-# Get version and architecture
+# Get version (prefer GITHUB_REF_NAME, then pyproject.toml, then git tag, then default)
 if [ -n "${GITHUB_REF_NAME:-}" ]; then
     VERSION="${GITHUB_REF_NAME#v}"
 else
-    VERSION=$(git describe --tags --abbrev=0 2>/dev/null | sed 's/^v//' || echo "3.0.4")
+    VERSION=$(
+        python3 - <<'PY'
+import sys, subprocess
+try:
+    import tomllib
+    with open("pyproject.toml","rb") as f:
+        cfg = tomllib.load(f)
+        v = cfg.get("project",{}).get("version")
+        if v:
+            print(v); sys.exit(0)
+except Exception:
+    pass
+try:
+    tag = subprocess.check_output(["git","describe","--tags","--abbrev=0"], stderr=subprocess.DEVNULL).decode().strip()
+    print(tag.lstrip("v")); sys.exit(0)
+except Exception:
+    print("3.0.4")
+PY
+    )
 fi
 
+# sanitize VERSION for filenames
+VERSION_SAFE="${VERSION//\//-}"
+VERSION_SAFE="${VERSION_SAFE// /-}"
+VERSION_SAFE="$(echo "$VERSION_SAFE" | tr -cd 'A-Za-z0-9._-')"
+
 ARCH=$(uname -m)
-echo "Version: ${VERSION}"
+echo "Version: ${VERSION} (safe: ${VERSION_SAFE})"
 echo "Architecture: ${ARCH}"
 
 # Verify single-file executable exists
@@ -82,7 +105,7 @@ if [ -f "LICENSE" ]; then
 fi
 
 # Create DMG
-DMG_NAME="pyprophet-${VERSION}-macos-${ARCH}.dmg"
+DMG_NAME="pyprophet-${VERSION_SAFE}-macos-${ARCH}.dmg"
 echo "Creating DMG: ${DMG_NAME}"
 
 # Remove any existing DMG