diff --git a/bonobo/constants.py b/bonobo/constants.py
index fceb8f9..b1a199c 100644
--- a/bonobo/constants.py
+++ b/bonobo/constants.py
@@ -11,9 +11,18 @@ class Token:
 BEGIN = Token('Begin')
 END = Token('End')
 
-INHERIT_INPUT = Token('InheritInput')
-LOOPBACK = Token('Loopback')
-NOT_MODIFIED = Token('NotModified')
+
+class Flag(Token):
+    must_be_first = False
+    must_be_last = False
+    allows_data = True
+
+
+INHERIT = Flag('Inherit')
+NOT_MODIFIED = Flag('NotModified')
+NOT_MODIFIED.must_be_first = True
+NOT_MODIFIED.must_be_last = True
+NOT_MODIFIED.allows_data = False
 
 EMPTY = tuple()
 
diff --git a/bonobo/examples/__init__.py b/bonobo/examples/__init__.py
index e1815f8..ec68fc5 100644
--- a/bonobo/examples/__init__.py
+++ b/bonobo/examples/__init__.py
@@ -27,6 +27,6 @@ def get_graph_options(options):
     _print = options.pop('print', False)
 
     return {
-        '_limit': (bonobo.Limit(_limit),) if _limit else (),
-        '_print': (bonobo.PrettyPrinter(),) if _print else (),
+        '_limit': (bonobo.Limit(_limit), ) if _limit else (),
+        '_print': (bonobo.PrettyPrinter(), ) if _print else (),
     }
diff --git a/bonobo/examples/datasets/__main__.py b/bonobo/examples/datasets/__main__.py
index 768ac5c..55a7fde 100644
--- a/bonobo/examples/datasets/__main__.py
+++ b/bonobo/examples/datasets/__main__.py
@@ -66,11 +66,16 @@ def get_services():
 
 if __name__ == '__main__':
     parser = examples.get_argument_parser()
-    parser.add_argument('--target', '-t', choices=graphs.keys(), nargs='+')
+    parser.add_argument(
+        '--target', '-t', choices=graphs.keys(), nargs='+'
+    )
 
     with bonobo.parse_args(parser) as options:
         graph_options = examples.get_graph_options(options)
-        graph_names = list(options['target'] if options['target'] else sorted(graphs.keys()))
+        graph_names = list(
+            options['target']
+            if options['target'] else sorted(graphs.keys())
+        )
 
         graph = bonobo.Graph()
         for name in graph_names:
diff --git a/bonobo/execution/contexts/__init__.py b/bonobo/execution/contexts/__init__.py
index e69de29..4c462c5 100644
--- a/bonobo/execution/contexts/__init__.py
+++ b/bonobo/execution/contexts/__init__.py
@@ -0,0 +1,9 @@
+from bonobo.execution.contexts.graph import GraphExecutionContext
+from bonobo.execution.contexts.node import NodeExecutionContext
+from bonobo.execution.contexts.plugin import PluginExecutionContext
+
+__all__ = [
+    'GraphExecutionContext',
+    'NodeExecutionContext',
+    'PluginExecutionContext',
+]
diff --git a/bonobo/execution/contexts/node.py b/bonobo/execution/contexts/node.py
index 194cf36..316d9b8 100644
--- a/bonobo/execution/contexts/node.py
+++ b/bonobo/execution/contexts/node.py
@@ -7,11 +7,11 @@ from types import GeneratorType
 
 from bonobo.config import create_container
 from bonobo.config.processors import ContextCurrifier
-from bonobo.constants import NOT_MODIFIED, BEGIN, END, TICK_PERIOD, Token
+from bonobo.constants import NOT_MODIFIED, BEGIN, END, TICK_PERIOD, Token, Flag, INHERIT
 from bonobo.errors import InactiveReadableError, UnrecoverableError, UnrecoverableTypeError
 from bonobo.execution.contexts.base import BaseContext
 from bonobo.structs.inputs import Input
-from bonobo.util import get_name, istuple, isconfigurabletype, ensure_tuple
+from bonobo.util import get_name, isconfigurabletype, ensure_tuple
 from bonobo.util.bags import BagType
 from bonobo.util.statistics import WithStatistics
 
@@ -292,20 +292,24 @@ class NodeExecutionContext(BaseContext, WithStatistics):
 
     def _cast(self, _input, _output):
         """
-        Transforms a pair of input/output into what is the real output.
+        Transforms a pair of input/output into the real slim output.
 
         :param _input: Bag
         :param _output: mixed
         :return: Bag
         """
 
-        if _output is NOT_MODIFIED:
-            if self._output_type is None:
-                return _input
-            else:
-                return self._output_type(*_input)
+        tokens, _output = split_token(_output)
 
-        return ensure_tuple(_output, cls=(self.output_type or tuple))
+        if NOT_MODIFIED in tokens:
+            return ensure_tuple(_input, cls=(self.output_type or tuple))
+
+        if INHERIT in tokens:
+            if self._output_type is None:
+                self._output_type = concat_types(self._input_type, self._input_length, self._output_type, len(_output))
+            _output = _input + ensure_tuple(_output)
+
+        return ensure_tuple(_output, cls=(self._output_type or tuple))
 
     def _send(self, value, _control=False):
         """
@@ -330,26 +334,44 @@ class NodeExecutionContext(BaseContext, WithStatistics):
 
 
 def isflag(param):
-    return isinstance(param, Token) and param in (NOT_MODIFIED, )
+    return isinstance(param, Flag)
 
 
-def split_tokens(output):
+def split_token(output):
     """
     Split an output into token tuple, real output tuple.
 
     :param output:
     :return: tuple, tuple
     """
-    if isinstance(output, Token):
-        # just a flag
-        return (output, ), ()
 
-    if not istuple(output):
-        # no flag
-        return (), (output, )
+    output = ensure_tuple(output)
 
-    i = 0
-    while isflag(output[i]):
+    flags, i, len_output, data_allowed = set(), 0, len(output), True
+    while i < len_output and isflag(output[i]):
+        if output[i].must_be_first and i:
+            raise ValueError('{} flag must be first.'.format(output[i]))
+        if i and output[i - 1].must_be_last:
+            raise ValueError('{} flag must be last.'.format(output[i - 1]))
+        if output[i] in flags:
+            raise ValueError('Duplicate flag {}.'.format(output[i]))
+        flags.add(output[i])
+        data_allowed &= output[i].allows_data
         i += 1
 
-    return output[:i], output[i:]
+    output = output[i:]
+    if not data_allowed and len(output):
+        raise ValueError('Output data provided after a flag that does not allow data.')
+    return flags, output
+
+
+def concat_types(t1, l1, t2, l2):
+    t1, t2 = t1 or tuple, t2 or tuple
+
+    if t1 == t2 == tuple:
+        return tuple
+
+    f1 = t1._fields if hasattr(t1, '_fields') else tuple(range(l1))
+    f2 = t2._fields if hasattr(t2, '_fields') else tuple(range(l2))
+
+    return BagType('Inherited', f1 + f2)
diff --git a/bonobo/nodes/basics.py b/bonobo/nodes/basics.py
index 9710ef7..ecf6827 100644
--- a/bonobo/nodes/basics.py
+++ b/bonobo/nodes/basics.py
@@ -1,11 +1,7 @@
 import functools
+import html
 import itertools
-import operator
 import pprint
-from functools import reduce
-
-from bonobo.util import ensure_tuple
-from mondrian import term
 
 from bonobo import settings
 from bonobo.config import Configurable, Option, Method, use_raw_input, use_context, use_no_input
@@ -14,6 +10,7 @@ from bonobo.config.processors import ContextProcessor, use_context_processor
 from bonobo.constants import NOT_MODIFIED
 from bonobo.util.objects import ValueHolder
 from bonobo.util.term import CLEAR_EOL
+from mondrian import term
 
 __all__ = [
     'FixedWindow',
@@ -94,29 +91,41 @@ class PrettyPrinter(Configurable):
 
     @ContextProcessor
     def context(self, context):
+        context.setdefault('_jupyter_html', None)
         yield context
+        if context._jupyter_html is not None:
+            from IPython.display import display, HTML
+            display(HTML('\n'.join(['<table>'] + context._jupyter_html + ['</table>'])))
 
     def __call__(self, context, *args, **kwargs):
-        quiet = settings.QUIET.get()
-        formater = self._format_quiet if quiet else self._format_console
-
-        if not quiet:
-            print('\u250e' + '\u2500' * (self.max_width - 1))
-
-        for index, (key, value) in enumerate(itertools.chain(enumerate(args), kwargs.items())):
-            if self.filter(index, key, value):
-                print(formater(index, key, value, fields=context.get_input_fields()))
-
-        if not quiet:
-            print('\u2516' + '\u2500' * (self.max_width - 1))
+        if not settings.QUIET:
+            if term.isjupyter:
+                self.print_jupyter(context, *args, **kwargs)
+                return NOT_MODIFIED
+            if term.istty:
+                self.print_console(context, *args, **kwargs)
+                return NOT_MODIFIED
 
+        self.print_quiet(context, *args, **kwargs)
         return NOT_MODIFIED
 
-    def _format_quiet(self, index, key, value, *, fields=None):
+    def print_quiet(self, context, *args, **kwargs):
+        for index, (key, value) in enumerate(itertools.chain(enumerate(args), kwargs.items())):
+            if self.filter(index, key, value):
+                print(self.format_quiet(index, key, value, fields=context.get_input_fields()))
+
+    def format_quiet(self, index, key, value, *, fields=None):
         # XXX should we implement argnames here ?
         return ' '.join(((' ' if index else '-'), str(key), ':', str(value).strip()))
 
-    def _format_console(self, index, key, value, *, fields=None):
+    def print_console(self, context, *args, **kwargs):
+        print('\u250e' + '\u2500' * (self.max_width - 1))
+        for index, (key, value) in enumerate(itertools.chain(enumerate(args), kwargs.items())):
+            if self.filter(index, key, value):
+                print(self.format_console(index, key, value, fields=context.get_input_fields()))
+        print('\u2516' + '\u2500' * (self.max_width - 1))
+
+    def format_console(self, index, key, value, *, fields=None):
         fields = fields or []
         if not isinstance(key, str):
             if len(fields) >= key and str(key) != str(fields[key]):
@@ -136,6 +145,21 @@ class PrettyPrinter(Configurable):
         ).strip()
         return '{}{}{}'.format(prefix, repr_of_value.replace('\n', CLEAR_EOL + '\n'), CLEAR_EOL)
 
+    def print_jupyter(self, context, *args):
+        if not context._jupyter_html:
+            context._jupyter_html = [
+                '<thead><tr>',
+                *map('<th>{}</th>'.format, map(html.escape, map(str,
+                                                                context.get_input_fields() or range(len(args))))),
+                '</tr></thead>',
+            ]
+
+        context._jupyter_html += [
+            '<tr>',
+            *map('<td>{}</td>'.format, map(html.escape, map(repr, args))),
+            '</tr>',
+        ]
+
 
 @use_no_input
 def noop(*args, **kwargs):
diff --git a/bonobo/settings.py b/bonobo/settings.py
index fdc4412..799ba3d 100644
--- a/bonobo/settings.py
+++ b/bonobo/settings.py
@@ -46,6 +46,9 @@ class Setting:
     def __eq__(self, other):
         return self.get() == other
 
+    def __bool__(self):
+        return bool(self.get())
+
     def set(self, value):
         value = self.formatter(value) if self.formatter else value
         if self.validator and not self.validator(value):
diff --git a/docs/conf.py b/docs/conf.py
index 46a9b55..53ed816 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -192,6 +192,4 @@ rst_epilog = """
 
 .. |longversion| replace:: v.{version}
 
-""".format(
-    version=version,
-)
+""".format(version=version, )
diff --git a/tests/execution/contexts/test_node.py b/tests/execution/contexts/test_node.py
index 0ebae6d..c9c8f1f 100644
--- a/tests/execution/contexts/test_node.py
+++ b/tests/execution/contexts/test_node.py
@@ -3,8 +3,8 @@ from unittest.mock import MagicMock
 import pytest
 
 from bonobo import Graph
-from bonobo.constants import EMPTY
-from bonobo.execution.contexts.node import NodeExecutionContext
+from bonobo.constants import EMPTY, NOT_MODIFIED, INHERIT
+from bonobo.execution.contexts.node import NodeExecutionContext, split_token
 from bonobo.execution.strategies import NaiveStrategy
 from bonobo.util.testing import BufferingNodeExecutionContext, BufferingGraphExecutionContext
 
@@ -224,3 +224,35 @@ def test_node_lifecycle_with_kill():
 
     ctx.stop()
     assert all((ctx.started, ctx.killed, ctx.stopped)) and not ctx.alive
+
+
+def test_split_token():
+    assert split_token(('foo', 'bar')) == (set(), ('foo', 'bar'))
+    assert split_token(()) == (set(), ())
+    assert split_token('') == (set(), ('', ))
+
+
+def test_split_token_duplicate():
+    with pytest.raises(ValueError):
+        split_token((NOT_MODIFIED, NOT_MODIFIED))
+    with pytest.raises(ValueError):
+        split_token((INHERIT, INHERIT))
+    with pytest.raises(ValueError):
+        split_token((INHERIT, NOT_MODIFIED, INHERIT))
+
+
+def test_split_token_not_modified():
+    with pytest.raises(ValueError):
+        split_token((NOT_MODIFIED, 'foo', 'bar'))
+    with pytest.raises(ValueError):
+        split_token((NOT_MODIFIED, INHERIT))
+    with pytest.raises(ValueError):
+        split_token((INHERIT, NOT_MODIFIED))
+    assert split_token(NOT_MODIFIED) == ({NOT_MODIFIED}, ())
+    assert split_token((NOT_MODIFIED, )) == ({NOT_MODIFIED}, ())
+
+
+def test_split_token_inherit():
+    assert split_token(INHERIT) == ({INHERIT}, ())
+    assert split_token((INHERIT, )) == ({INHERIT}, ())
+    assert split_token((INHERIT, 'foo', 'bar')) == ({INHERIT}, ('foo', 'bar'))
diff --git a/tests/features/test_inherit.py b/tests/features/test_inherit.py
new file mode 100644
index 0000000..92b943b
--- /dev/null
+++ b/tests/features/test_inherit.py
@@ -0,0 +1,27 @@
+from bonobo.constants import INHERIT
+from bonobo.util.testing import BufferingNodeExecutionContext
+
+messages = [
+    ('Hello', ),
+    ('Goodbye', ),
+]
+
+
+def append(*args):
+    return INHERIT, '!'
+
+
+def test_inherit():
+    with BufferingNodeExecutionContext(append) as context:
+        context.write_sync(*messages)
+
+    assert context.get_buffer() == list(map(lambda x: x + ('!', ), messages))
+
+
+def test_inherit_bag_tuple():
+    with BufferingNodeExecutionContext(append) as context:
+        context.set_input_fields(['message'])
+        context.write_sync(*messages)
+
+    assert context.get_output_fields() == ('message', '0')
+    assert context.get_buffer() == list(map(lambda x: x + ('!', ), messages))