diff --git a/inspirehep/modules/workflows/errors.py b/inspirehep/modules/workflows/errors.py
index b5ecf21865..7e706df524 100644
--- a/inspirehep/modules/workflows/errors.py
+++ b/inspirehep/modules/workflows/errors.py
@@ -26,5 +26,11 @@
 
 
 class DownloadError(WorkflowsError):
-
     """Error representing a failed download in a workflow."""
+
+
+class MissingHeadUUIDError(WorkflowsError):
+    """Error representing the missing field `head_uuid`
+
+    This error should be triggered when there is no head_uuid property
+    inside the extra_data for a given workflow object."""
diff --git a/inspirehep/modules/workflows/tasks/merging.py b/inspirehep/modules/workflows/tasks/merging.py
new file mode 100644
index 0000000000..e3cb5f6a85
--- /dev/null
+++ b/inspirehep/modules/workflows/tasks/merging.py
@@ -0,0 +1,192 @@
+# -*- coding: utf-8 -*-
+#
+# This file is part of INSPIRE.
+# Copyright (C) 2014-2017 CERN.
+#
+# INSPIRE is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# INSPIRE is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with INSPIRE. If not, see <http://www.gnu.org/licenses/>.
+#
+# In applying this license, CERN does not waive the privileges and immunities
+# granted to it by virtue of its status as an Intergovernmental Organization
+# or submit itself to any jurisdiction.
+
+"""Tasks related to record merging."""
+
+from __future__ import absolute_import, division, print_function
+
+from invenio_db import db
+from invenio_pidstore.resolver import PersistentIdentifier
+from inspire_json_merger.inspire_json_merger import inspire_json_merge
+
+from inspirehep.modules.records import RecordMetadata
+from inspirehep.modules.records.api import InspireRecord
+from inspirehep.utils.record import get_source
+
+from inspirehep.modules.workflows.models import WorkflowsRecordSources
+
+from inspirehep.modules.workflows.errors import MissingHeadUUIDError
+from inspirehep.modules.workflows.utils import with_debug_logging
+
+
+def put_matched_record_uuid_in_extra_data(obj):
+    """Retrieve the UUID from the JSON record.
+
+    Retrieves the record UUID .
+    The uuid is stored in `obj.extra_data['head_uuid']`.
+    """
+    try:
+        record_matched = obj.extra_data['record_matches']['records'][0]
+        match_id = record_matched['source']['control_number']
+    except Exception as e:
+        raise ValueError(
+            'Can not get control number from '
+            'matched record.\n{}'.format(e.message)
+        )
+    record_uuid = PersistentIdentifier.get('lit', match_id).object_uuid
+    obj.extra_data['head_uuid'] = str(record_uuid)
+
+
+def get_head(obj):
+    """Retrieve the head from the JSON record.
+
+    Retrieves the head JSON corresponding to the current matched record.
+    The head version is stored in `obj.extra_data['head_json']`.
+    It assumes that inside obj.extra_data there is head_uuid key populated
+    """
+    head_uuid = obj.extra_data.get('head_uuid')
+    if not head_uuid:
+        raise MissingHeadUUIDError
+
+    entry = RecordMetadata.query.filter(
+        RecordMetadata.id == head_uuid
+    ).one_or_none()
+
+    return entry.json if entry else {}
+
+
+@with_debug_logging
+def merge_articles(obj, eng):
+    """Retrieve root, head, update and perform the merge.
+
+    - The workflow payload is overwritten by the merged record.
+
+    - The conflicts are stored in obj.extra_data['conflicts']. This variable
+        contains None if there aren't conflicts, otherwise a string made by
+        dumping a dictionary.
+    """
+    root = get_root(obj)
+    put_matched_record_uuid_in_extra_data(obj)
+    head = get_head(obj)
+    obj.data, conflicts = inspire_json_merge(
+        root,
+        head,
+        obj.extra_data['new_root']
+    )
+    obj.extra_data['conflicts'] = conflicts if conflicts else None
+
+
+def put_root_in_extra_data(obj, eng):
+    """Save the workflow object payload in extra_data['new_root']
+    to make it available later.
+    """
+    if not obj.extra_data.get('new_root'):
+        obj.extra_data['new_root'] = obj.data
+
+
+def update_record(obj, eng):
+    """Stores the merged record in the database.
+
+    When this function is called, it assumes:
+    - obj.extra_data['head_uuid'] is populated
+    - obj.data is populated with the json merged
+    """
+    record = InspireRecord.get_record(obj.extra_data['head_uuid'])
+    record.clear()
+    record.update(obj.data)
+    record.commit()
+    obj.save()
+    db.session.commit()
+
+
+def get_root(obj):
+    """Retrieve the root JSON.
+
+    Retrieves the root JSON corresponding to the current matched record and the
+    provided source.
+    The root version is stored in `obj.extra_data['root_json']`.
+
+    Args:
+        obj(WorkflowObject): the current workflow object
+
+    Return:
+        (dict): the json of the entry if it exists, else an empty dict
+    """
+    source = get_source(obj.data)
+    control_number = obj.data.get('control_number')
+    record_uuid = PersistentIdentifier.get('lit', control_number).object_uuid
+
+    return _read_wf_record_source(record_uuid, source)
+
+
+def _read_wf_record_source(record_uuid, source):
+    entry = WorkflowsRecordSources.query.filter(
+        WorkflowsRecordSources.record_id == record_uuid,
+        WorkflowsRecordSources.source == source
+    ).one_or_none()
+    return entry.json if entry else {}
+
+
+@with_debug_logging
+def store_root(obj, eng):
+    """Stores the root record in the database.
+
+    When this function is called, it assumes:
+    - obj.extra_data['head_uuid'] is populated
+    - obj.data is populated with the json merged
+    """
+    new_root = obj.extra_data['new_root']
+    record_uuid = obj.extra_data['head_uuid']
+    source = get_source(new_root)
+
+    _insert_wf_record_source(new_root, record_uuid, source)
+
+    # this line prevent emptying obj.extra_data
+    obj.save()
+    # Commit to DB before indexing
+    db.session.commit()
+
+
+def _insert_wf_record_source(json, record_uuid, source):
+    """
+    Stores the given json in the WorkflowRecordSource table in the db
+
+    Important: does not commit the session, in case some other operation
+    needs to be done before it
+    """
+    record_source = WorkflowsRecordSources.query.filter(
+        WorkflowsRecordSources.record_id == record_uuid,
+        WorkflowsRecordSources.source == source
+    ).one_or_none()
+    if record_source is None:
+        record_source = WorkflowsRecordSources(
+            source=source,
+            json=json,
+            record_id=record_uuid
+        )
+    else:
+        record_source.json = json
+    db.session.add(record_source)
+
+
+def has_conflicts(obj, eng):
+    return obj.extra_data.get('conflicts') is not None
diff --git a/inspirehep/modules/workflows/tasks/upload.py b/inspirehep/modules/workflows/tasks/upload.py
index 5d2cbaeb49..3454377f26 100644
--- a/inspirehep/modules/workflows/tasks/upload.py
+++ b/inspirehep/modules/workflows/tasks/upload.py
@@ -52,11 +52,15 @@ def store_record(obj, *args, **kwargs):
     # Create persistent identifier.
     inspire_recid_minter(str(record.id), record)
 
+    # store head_uuid to store the root later
+    obj.extra_data['head_uuid'] = str(record.id)
+
     # Commit any changes to record
     record.commit()
 
     # Dump any changes to record
     obj.data = record.dumps()
+    obj.save()
 
     # Commit to DB before indexing
     db.session.commit()
diff --git a/inspirehep/modules/workflows/utils.py b/inspirehep/modules/workflows/utils.py
index 63a7b4884f..1873f60334 100644
--- a/inspirehep/modules/workflows/utils.py
+++ b/inspirehep/modules/workflows/utils.py
@@ -36,7 +36,6 @@
 
 from .models import WorkflowsAudit
 
-
 LOGGER = logging.getLogger(__name__)
 
 
diff --git a/inspirehep/utils/record.py b/inspirehep/utils/record.py
index 54d3edb1f1..8da86fedbf 100644
--- a/inspirehep/utils/record.py
+++ b/inspirehep/utils/record.py
@@ -108,3 +108,8 @@ def is_complete(publication_info):
                 if is_complete(publication_info):
                     return False
     return had_at_least_one_journal_title
+
+
+def get_source(record):
+    """Return the first available source of a record."""
+    return get_value(record, 'acquisition_source.source')
diff --git a/tests/integration/workflows/test_workflows_merging.py b/tests/integration/workflows/test_workflows_merging.py
new file mode 100644
index 0000000000..1caf8d534c
--- /dev/null
+++ b/tests/integration/workflows/test_workflows_merging.py
@@ -0,0 +1,67 @@
+# -*- coding: utf-8 -*-
+#
+# This file is part of INSPIRE.
+# Copyright (C) 2014-2017 CERN.
+#
+# INSPIRE is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# INSPIRE is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with INSPIRE. If not, see <http://www.gnu.org/licenses/>.
+#
+# In applying this license, CERN does not waive the privileges and immunities
+# granted to it by virtue of its status as an Intergovernmental Organization
+# or submit itself to any jurisdiction.
+
+from __future__ import absolute_import, division, print_function
+
+import pytest
+
+from invenio_db import db
+
+from inspirehep.modules.records.api import InspireRecord
+from inspirehep.modules.workflows.tasks.merging import (
+    _read_wf_record_source,
+    _insert_wf_record_source
+)
+
+
+@pytest.fixture()
+def dummy_record(workflow_app):
+    record = InspireRecord.create({
+        "$schema": "http://localhost:5000/schemas/records/hep.json",
+        "titles": [{
+            "title": "foo"
+        }],
+        "document_type": ["thesis"]
+    })
+    yield record
+    record._delete(force=True)
+
+
+def test_wf_record_source_read_and_write(dummy_record):
+    record_uuid = dummy_record.id
+    original_root = {"title": "baz"}
+
+    _insert_wf_record_source(
+        json=original_root,
+        record_uuid=record_uuid,
+        source='arXiv'
+    )
+    db.session.commit()
+
+    retrieved_root = _read_wf_record_source(record_uuid=record_uuid, source='arXiv')
+    assert original_root == retrieved_root
+
+
+def test_empty_root(dummy_record):
+    record_uuid = dummy_record.id
+    retrieved_root = _read_wf_record_source(record_uuid=record_uuid, source='Elsevier')
+    assert retrieved_root == {}
diff --git a/tests/unit/workflows/test_workflows_tasks_merging.py b/tests/unit/workflows/test_workflows_tasks_merging.py
new file mode 100644
index 0000000000..b7ea078416
--- /dev/null
+++ b/tests/unit/workflows/test_workflows_tasks_merging.py
@@ -0,0 +1,44 @@
+# -*- coding: utf-8 -*-
+#
+# This file is part of INSPIRE.
+# Copyright (C) 2014-2017 CERN.
+#
+# INSPIRE is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# INSPIRE is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with INSPIRE. If not, see <http://www.gnu.org/licenses/>.
+#
+# In applying this license, CERN does not waive the privileges and immunities
+# granted to it by virtue of its status as an Intergovernmental Organization
+# or submit itself to any jurisdiction.
+
+from __future__ import absolute_import, division, print_function
+
+from mocks import MockObj
+import pytest
+
+
+from inspirehep.modules.workflows.tasks.merging import get_head
+from inspirehep.modules.workflows.errors import MissingHeadUUIDError
+
+
+def test_get_head_empty_head_uuid():
+    workflow_obj = MockObj({}, {'head_uuid': ''})
+
+    with pytest.raises(MissingHeadUUIDError):
+        get_head(workflow_obj)
+
+
+def test_get_head_no_head_uuid():
+    workflow_obj = MockObj({}, {})
+
+    with pytest.raises(MissingHeadUUIDError):
+        get_head(workflow_obj)