diff --git a/src/matchbox/client/locations.py b/src/matchbox/client/locations.py
index dcd28931..11b3c9b2 100644
--- a/src/matchbox/client/locations.py
+++ b/src/matchbox/client/locations.py
@@ -152,6 +152,7 @@ def execute(
         rename: dict[str, str] | Callable | None = None,
         return_type: QueryReturnType = QueryReturnType.POLARS,
         keys: tuple[str, list[str]] | None = None,
+        limit: int | None = None,
     ) -> Iterator[QueryReturnClass]:
         """Execute ET logic against location and return batches.
 
@@ -168,6 +169,7 @@ def execute(
             keys: Rule to only retrieve rows by specific keys.
                 The key of the dictionary is a field name on which to filter.
                 Filters source entries where the key field is in the dict values.
+            limit: Maximum number of rows to return. If None, returns all rows.
 
         Raises:
             AttributeError: If the cliet is not set.
@@ -328,6 +330,7 @@ def execute(
         return_type: Literal[QueryReturnType.POLARS] = ...,
         keys: tuple[str, list[str]] | None = None,
         schema_overrides: dict[str, pl.DataType] | None = None,
+        limit: int | None = None,
     ) -> Generator[PolarsDataFrame, None, None]: ...
 
     @overload
@@ -339,6 +342,7 @@ def execute(
         return_type: Literal[QueryReturnType.PANDAS] = ...,
         keys: tuple[str, list[str]] | None = None,
         schema_overrides: dict[str, pl.DataType] | None = None,
+        limit: int | None = None,
     ) -> Generator[PandasDataFrame, None, None]: ...
 
     @overload
@@ -350,6 +354,7 @@ def execute(
         return_type: Literal[QueryReturnType.ARROW] = ...,
         keys: tuple[str, list[str]] | None = None,
         schema_overrides: dict[str, pl.DataType] | None = None,
+        limit: int | None = None,
     ) -> Generator[ArrowTable, None, None]: ...
 
     @requires_client
@@ -361,6 +366,7 @@ def execute(  # noqa: D102
         return_type: QueryReturnType = QueryReturnType.POLARS,
         keys: tuple[str, list[str]] | None = None,
         schema_overrides: dict[str, pl.DataType] | None = None,
+        limit: int | None = None,
     ) -> Generator[QueryReturnClass, None, None]:
         # Strip semicolon, as it can block extended query protocol
         # and slow some engines down
@@ -382,6 +388,11 @@ def execute(  # noqa: D102
                         f"select * from ({extract_transform}) as sub "
                         f"where {key_field} in ({comma_separated_values})"
                     )
+            # Add LIMIT clause if limit is specified
+            if limit is not None:
+                extract_transform = (
+                    f"select * from ({extract_transform}) as sub_limit limit {limit}"
+                )
             yield from sql_to_df(
                 stmt=extract_transform,
                 schema_overrides=schema_overrides,
diff --git a/src/matchbox/client/sources.py b/src/matchbox/client/sources.py
index 4b4656c0..4a0dc68c 100644
--- a/src/matchbox/client/sources.py
+++ b/src/matchbox/client/sources.py
@@ -239,6 +239,7 @@ def fetch(
         batch_size: int | None = None,
         return_type: Literal[QueryReturnType.POLARS] = ...,
         keys: list[str] | None = None,
+        limit: int | None = None,
     ) -> Generator[PolarsDataFrame, None, None]: ...
 
     @overload
@@ -248,6 +249,7 @@ def fetch(
         batch_size: int | None = None,
         return_type: Literal[QueryReturnType.PANDAS] = ...,
         keys: list[str] | None = None,
+        limit: int | None = None,
     ) -> Generator[PandasDataFrame, None, None]: ...
 
     @overload
@@ -257,6 +259,7 @@ def fetch(
         batch_size: int | None = None,
         return_type: Literal[QueryReturnType.ARROW] = ...,
         keys: list[str] | None = None,
+        limit: int | None = None,
     ) -> Generator[ArrowTable, None, None]: ...
 
     def fetch(
@@ -265,6 +268,7 @@ def fetch(
         batch_size: int | None = None,
         return_type: QueryReturnType = QueryReturnType.POLARS,
         keys: list[str] | None = None,
+        limit: int | None = None,
     ) -> Generator[QueryReturnClass, None, None]:
         """Applies the extract/transform logic to the source and returns batches lazily.
 
@@ -274,6 +278,7 @@ def fetch(
             batch_size: Indicate the size of each batch when fetching data in batches.
             return_type: The type of data to return. Defaults to "polars".
             keys: List of keys to select a subset of all source entries.
+            limit: Maximum number of rows to return. If None, returns all rows.
 
         Returns:
             The requested data in the specified format, as an iterator of tables.
@@ -295,6 +300,7 @@ def _rename(c: str) -> str:
                 batch_size=batch_size,
                 return_type=return_type,
                 keys=(self.config.key_field.name, keys),
+                limit=limit,
             )
         else:
             yield from self.location.execute(
@@ -303,13 +309,30 @@ def _rename(c: str) -> str:
                 rename=_rename,
                 batch_size=batch_size,
                 return_type=return_type,
+                limit=limit,
             )
 
     def sample(
         self, n: int = 100, return_type: QueryReturnType = QueryReturnType.POLARS
-    ) -> None:
-        """Peek at the top n entries in a source."""
-        return next(self.fetch(batch_size=n, return_type=return_type))
+    ) -> QueryReturnClass:
+        """Peek at the top n entries in a source.
+
+        Args:
+            n: The number of rows to return. Must be a positive integer.
+            return_type: The type of data to return. Defaults to "polars".
+
+        Returns:
+            A dataframe containing the first n rows of the source.
+
+        Raises:
+            TypeError: If n is not an integer.
+            ValueError: If n is not a positive integer.
+        """
+        if not isinstance(n, int):
+            raise TypeError(f"n must be an integer, got {type(n).__name__}")
+        if n <= 0:
+            raise ValueError(f"n must be a positive integer, got {n}")
+        return next(self.fetch(limit=n, return_type=return_type))
 
     @profile_time(attr="name")
     def run(self, batch_size: int | None = None) -> ArrowTable:
diff --git a/test/client/test_locations.py b/test/client/test_locations.py
index a2e191a6..07e57c12 100644
--- a/test/client/test_locations.py
+++ b/test/client/test_locations.py
@@ -303,6 +303,10 @@ def test_relational_db_execute(
     unfiltered_results = pl.concat(location.execute(sql, batch_size, keys=("key", [])))
     assert_frame_equal(unfiltered_results, combined_df)
 
+    # Try using limit to restrict results
+    limited_results = pl.concat(location.execute(sql, batch_size, limit=3))
+    assert len(limited_results) == 3
+
 
 @pytest.mark.parametrize(
     "warehouse",
diff --git a/test/client/test_sources.py b/test/client/test_sources.py
index 0e02a5fa..d6ee4b56 100644
--- a/test/client/test_sources.py
+++ b/test/client/test_sources.py
@@ -155,6 +155,51 @@ def test_source_fetch_and_sample(sqla_sqlite_warehouse: Engine) -> None:
     assert len(result) == 5
 
 
+def test_source_sample_validation(sqla_sqlite_warehouse: Engine) -> None:
+    """Test that source.sample validates input correctly."""
+    # Create test data
+    source_testkit = source_factory(
+        n_true_entities=5,
+        features=[
+            {"name": "name", "base_generator": "word", "datatype": DataTypes.STRING},
+        ],
+        engine=sqla_sqlite_warehouse,
+    ).write_to_location()
+
+    # Create location and source
+    location = RelationalDBLocation(name="dbname").set_client(sqla_sqlite_warehouse)
+    source = Source(
+        dag=source_testkit.source.dag,
+        location=location,
+        name="test_source",
+        extract_transform=source_testkit.source_config.extract_transform,
+        infer_types=True,
+        key_field="key",
+        index_fields=["name"],
+    )
+
+    # Test that string n raises TypeError
+    with pytest.raises(TypeError, match="n must be an integer"):
+        source.sample(n="asdf")
+
+    # Test that negative n raises ValueError
+    with pytest.raises(ValueError, match="n must be a positive integer"):
+        source.sample(n=-1)
+
+    # Test that zero n raises ValueError
+    with pytest.raises(ValueError, match="n must be a positive integer"):
+        source.sample(n=0)
+
+    # Test that float n raises TypeError
+    with pytest.raises(TypeError, match="n must be an integer"):
+        source.sample(n=10.5)
+
+    # Test that valid positive integer works
+    result = source.sample(n=3)
+    assert isinstance(result, pl.DataFrame)
+    assert len(result) == 3
+
+
 @pytest.mark.parametrize(
     "qualify_names",
     [