diff --git a/setup.py b/setup.py
--- a/setup.py
+++ b/setup.py
@@ -54,6 +54,8 @@
     entry_points="""
         [swh.workers]
         loader.mercurial=swh.loader.mercurial:register
+        [console_scripts]
+        swh-hg-identify=swh.loader.mercurial.identify:main
     """,
     classifiers=[
         "Programming Language :: Python :: 3",
diff --git a/swh/loader/mercurial/identify.py b/swh/loader/mercurial/identify.py
new file mode 100644
--- /dev/null
+++ b/swh/loader/mercurial/identify.py
@@ -0,0 +1,532 @@
+# Copyright (C) 2018-2020  The Software Heritage developers
+# See the AUTHORS file at the top-level directory of this distribution
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+import json
+import re
+import subprocess
+from codecs import escape_decode  # type: ignore
+from pathlib import Path
+from typing import Any, Dict, Iterator, List, NamedTuple, Optional, Union
+
+# WARNING: do not import unnecessary things here to keep cli startup time under
+# control
+import click
+
+from swh.model.hashutil import hash_to_bytehex
+from swh.model.identifiers import normalize_timestamp, swhid
+from swh.model.model import RevisionType
+
+TAG_PATTERN = re.compile(b"([0-9A-Fa-f]{40}) +(.+)")
+
+
+class HgAuthor(NamedTuple):
+    """Represent a Mercurial revision author."""
+
+    fullname: bytes
+    """full name of the author"""
+
+    name: Optional[bytes]
+    """name of the author"""
+
+    email: Optional[bytes]
+    """email of the author"""
+
+    @staticmethod
+    def from_bytes(data: bytes) -> "HgAuthor":
+        """Convert bytes to an HgAuthor named tuple.
+
+        Expected format: "name <email>"
+        """
+        from swh.loader.mercurial.converters import parse_author
+
+        result = parse_author(data)
+        return HgAuthor(
+            fullname=result["fullname"], name=result["name"], email=result["email"]
+        )
+
+    def to_dict(self) -> Dict[str, Optional[bytes]]:
+        return {"fullname": self.fullname, "name": self.name, "email": self.email}
+
+
+HG_REVISION_TEMPLATE = "\n".join(
+    [
+        "node_id:{node}",
+        "author:{author}",
+        "timestamp_offset:{date|json}",
+        "p1:{p1.node}",
+        "p2:{p2.node}",
+        "extras:{join(extras, '\nextras:')}",
+    ]
+)  # Log template for HgRevision.from_bytes
+
+NULL_NODE_ID = b"0" * 40  # Value used when no parent
+
+
+class HgRevision(NamedTuple):
+    """Represent a Mercurial revision."""
+
+    node_id: bytes
+    """raw bytes of the revision hash"""
+
+    author: HgAuthor
+    """author of the revision"""
+
+    timestamp: bytes
+    """timestamp of the revision"""
+
+    offset: bytes
+    """offset of the revision"""
+
+    parents: List[bytes]
+    """hex bytes of the revision's parents"""
+
+    extras: Dict[bytes, bytes]
+    """metadata of the revision"""
+
+    description: bytes
+    """description of the revision"""
+
+    @staticmethod
+    def from_bytes(data: bytes, description: bytes) -> "HgRevision":
+        """Convert bytes to an HgRevision named tuple.
+
+        Expected data format:
+        '''
+        node_id:{node}
+        author:{author}
+        timestamp_offset:[{timestamp}, {offset}]
+        p1:{p1}
+        p2:{p2}
+        extras:{key1}={value1}
+        ...
+        extras:{keyn}={value}
+        '''
+
+        """
+        lines = data.split(b"\n")
+        tuples = [line.split(b":", 1) for line in lines]
+        fields: Dict[str, Any] = {
+            "parents": [],
+            "extras": {},
+            "description": description,
+        }
+        for key, value in tuples:
+            if key == b"timestamp_offset":
+                timestamp, offset = json.loads(value)
+                fields["timestamp"] = timestamp
+                fields["offset"] = offset
+            elif key in (b"p1", b"p2"):
+                if value != NULL_NODE_ID:
+                    fields["parents"].append(value)
+            elif key == b"extras":
+                extra_key, extra_value = value.split(b"=", 1)
+                fields["extras"][extra_key] = extra_value
+            elif key == b"author":
+                fields["author"] = HgAuthor.from_bytes(value)
+            else:
+                fields[key.decode()] = value
+
+        return HgRevision(**fields)
+
+    def branch(self) -> bytes:
+        return self.extras.get(b"branch", b"default")
+
+    def to_dict(self) -> Dict:
+        """Convert a HgRevision to a dict for SWHID computation"""
+        date = normalize_timestamp(int(self.timestamp))
+
+        extra_headers = [
+            (b"time_offset_seconds", str(self.offset).encode("utf-8")),
+        ]
+
+        for key, value in self.extras.items():
+            if key == b"branch" and value == b"default":
+                # branch default is skipped to match actual implementation
+                continue
+            if key == b"transplant_source":
+                # transplant_source is converted to hex to match actual implementation
+                value = hash_to_bytehex(escape_decode(value)[0])
+            extra_headers.append((key, value))
+
+        author = self.author.to_dict()
+
+        return {
+            "author": author,
+            "date": date,
+            "committer": author,
+            "committer_date": date,
+            "type": RevisionType.MERCURIAL.value,
+            "message": self.description,
+            "metadata": {"node": self.node_id},
+            "extra_headers": tuple(extra_headers),
+            "synthetic": False,
+            "parents": self.parents,
+        }
+
+
+class HgBranch(NamedTuple):
+    """Represent a Mercurial branch."""
+
+    name: bytes
+    """name of the branch"""
+
+    node_id: bytes
+    """row bytes of the target revision hash"""
+
+
+class HgTag(NamedTuple):
+    """Represent a Mercurial tag."""
+
+    name: bytes
+    """name of the tag"""
+
+    node_id: bytes
+    """hex bytes of the target revision"""
+
+
+class Hg:
+    """Provide methods to extract data from a Mercurial repository."""
+
+    def __init__(self, repository_root: Path) -> None:
+        self._root = repository_root
+
+    def _output(self, *args) -> bytes:
+        """Return the outpout of a `hg` call."""
+        return subprocess.check_output(["hg", *args], cwd=self._root)
+
+    def _call(self, *args) -> None:
+        """Perform a `hg` call."""
+        subprocess.check_call(
+            ["hg", *args],
+            cwd=self._root,
+            stderr=subprocess.PIPE,
+            stdout=subprocess.PIPE,
+        )
+
+    def root(self) -> Path:
+        """Return the root of the Mercurial repository."""
+        return self._root
+
+    def log(self, rev: Optional[Union[bytes, str]] = None) -> List[HgRevision]:
+        """Return the specified revisions of the Mercurial repository.
+
+        Mercurial revsets are supported. (See `hg help revsets`)
+
+        If no revision range is specified, return all revisions".
+        """
+        if rev:
+            node_ids = self._output("log", "-r", rev, "-T", "{node}\n").splitlines()
+        else:
+            node_ids = self._output("log", "-T", "{node}\n").splitlines()
+
+        revisions = [self._revision(node_id) for node_id in reversed(node_ids)]
+
+        return revisions
+
+    def _revision(self, revision: bytes) -> HgRevision:
+        data = self._output("log", "-r", revision, "-T", HG_REVISION_TEMPLATE)
+
+        # hg log strips the description so the raw description has to be taken
+        # from debugdata
+        _, desc = self._output("debugdata", "-c", revision).split(b"\n\n", 1)
+
+        return HgRevision.from_bytes(data, desc)
+
+    def up(self, rev: bytes) -> None:
+        """Update the repository working directory to the specified revision."""
+        self._call("up", rev)
+
+    def branches(self) -> List[HgBranch]:
+        """List the repository named branches."""
+        output = self._output("branches", "-T", "{branch}\n{node}\n\n").strip()
+
+        branches = []
+
+        for block in output.split(b"\n\n"):
+            name, node_id = block.splitlines()
+            branches.append(HgBranch(name=name, node_id=node_id))
+
+        return branches
+
+    def tip(self) -> HgRevision:
+        """Return the `tip` revision."""
+        return self.log("tip")[0]
+
+    def tags(self) -> List[HgTag]:
+        """Return the repository tags at the current revision."""
+        hgtags = self._root / ".hgtags"
+
+        tags = {}
+
+        if hgtags.is_file():
+            for line in hgtags.read_bytes().splitlines():
+                match = TAG_PATTERN.match(line)
+                if match is None:
+                    continue
+                node_id, name = match.groups()
+                tags[node_id] = name
+
+        return [HgTag(name=name, node_id=node_id) for node_id, name in tags.items()]
+
+
+@click.group()
+@click.option(
+    "--directory",
+    "-d",
+    help=("Path to the Mercurial repository. If unset, the current directory is used"),
+)
+@click.pass_context
+def main(ctx, directory=None):
+    """Compute the Software Heritage persistent identifier (SWHID) for the given
+       source code object(s).
+
+    For more details about SWHIDs see:
+
+    https://docs.softwareheritage.org/devel/swh-model/persistent-identifiers.html
+    """
+    # ensure that ctx.obj exists and is a dict (in case `cli()` is called
+    # by means other than the `if` block below)
+    ctx.ensure_object(dict)
+
+    root = Path(directory) if directory else Path()
+    if not root.exists():
+        raise IOError(f"{root!r} does not exists")
+
+    ctx.obj["HG_ROOT"] = root
+
+
+def identify_directory(path: Path) -> str:
+    """Return the SWHID of the given path."""
+    output = subprocess.check_output(
+        ["swh-identify", "--exclude", ".hg", "."], cwd=path
+    ).decode()
+    return output.split()[0].split(":")[-1]
+
+
+class RevisionIdentity(NamedTuple):
+    """Represent a swh revision identity."""
+
+    swhid: bytes
+    """SWHID raw bytes"""
+
+    node_id: bytes
+    """node_id hex bytes"""
+
+    directory_swhid: bytes
+
+    def dir_uri(self) -> str:
+        """Return the SWHID uri of the revision's directory."""
+        return f"swh:1:dir:{self.directory_swhid.hex()}\t{self.node_id.decode()}"
+
+    def __str__(self) -> str:
+        """Return the string representation of a RevisionIdentity."""
+        uri = swhid("revision", self.swhid.hex())
+        return f"{uri}\t{self.node_id.decode()}"
+
+
+def identify_revision(
+    hg: Hg,
+    rev: Optional[bytes] = None,
+    node_id_2_swhid: Optional[Dict[bytes, bytes]] = None,
+) -> Iterator[RevisionIdentity]:
+    """Return the repository revision identities.
+
+    hg: A `Hg` repository instance
+    rev: An optional revision or Mercurial revsets (See `hg help revsets`)
+         If not provided all the repository revisions will be computed.
+    node_id_2_swhid: An optional cache mapping hg node ids to SWHIDs
+        It will be updated in place with new mappings.
+    """
+    from swh.model.hashutil import hash_to_bytes
+    from swh.model.model import Revision
+
+    if node_id_2_swhid is None:
+        node_id_2_swhid = {}
+
+    for revision in hg.log(rev):
+        data = revision.to_dict()
+
+        hg.up(revision.node_id)
+        directory_swhid = hash_to_bytes(identify_directory(hg.root()))
+        data["directory"] = directory_swhid
+
+        parents = []
+        for parent in data["parents"]:
+            if parent not in node_id_2_swhid:
+                parent_revision = next(identify_revision(hg, parent, node_id_2_swhid))
+                node_id_2_swhid[parent] = parent_revision.swhid
+            parents.append(node_id_2_swhid[parent])
+        data["parents"] = parents
+
+        revision_swhid = hash_to_bytes(Revision.from_dict(data).id)
+        node_id_2_swhid[revision.node_id] = revision_swhid
+
+        yield RevisionIdentity(
+            swhid=revision_swhid,
+            node_id=revision.node_id,
+            directory_swhid=directory_swhid,
+        )
+
+
+class ReleaseIdentity(NamedTuple):
+    """Represent a swh release identity."""
+
+    swhid: str
+    """SWHID hex string"""
+
+    node_id: bytes
+    """node_id hex bytes"""
+
+    name: bytes
+    """name of the release"""
+
+    def __str__(self) -> str:
+        """Return the string representation of a ReleaseIdentity."""
+        uri = swhid("release", self.swhid)
+        return f"{uri}\t{self.name.decode()}"
+
+
+def identify_release(
+    hg: Hg, node_id_2_swhid: Optional[Dict[bytes, bytes]] = None,
+) -> Iterator[ReleaseIdentity]:
+    """Return the repository's release identities.
+
+    hg: A `Hg` repository instance
+    node_id_2_swhid: An optional cache mapping hg node ids to SWHIDs
+        If not provided it will be computed using `identify_revision`.
+    """
+    from swh.model.model import ObjectType, Release
+
+    if node_id_2_swhid is None:
+        node_id_2_swhid = {
+            revision.node_id: revision.swhid for revision in identify_revision(hg)
+        }
+
+    for tag in hg.tags():
+        data = {
+            "name": tag.name,
+            "target": node_id_2_swhid[tag.node_id],
+            "target_type": ObjectType.REVISION.value,
+            "message": None,
+            "metadata": None,
+            "synthetic": False,
+            "author": {"name": None, "email": None, "fullname": b""},
+            "date": None,
+        }
+
+        release_swhid = Release.from_dict(data).id
+
+        yield ReleaseIdentity(
+            swhid=release_swhid, node_id=tag.node_id, name=tag.name,
+        )
+
+
+def identify_snapshot(
+    hg: Hg,
+    node_id_2_swhid: Optional[Dict[bytes, bytes]] = None,
+    releases: Optional[List[ReleaseIdentity]] = None,
+) -> str:
+    """Return the repository snapshot identity.
+
+    hg: A `Hg` repository instance
+    node_id_2_swhid: An optional cache mapping hg node ids to SWHIDs
+         If not provided it will be computed using `identify_revision`.
+    release: an optional list of `ReleaseIdentity`.
+        If not provided it will be computed using `identify_release`.
+    """
+    from swh.model.model import Snapshot, TargetType
+
+    if node_id_2_swhid is None:
+        node_id_2_swhid = {
+            revision.node_id: revision.swhid for revision in identify_revision(hg)
+        }
+
+    if releases is None:
+        releases = [release for release in identify_release(hg, node_id_2_swhid)]
+
+    branches = {}
+
+    tip = hg.tip()
+    branches[b"HEAD"] = {
+        "target": tip.branch(),
+        "target_type": TargetType.ALIAS.value,
+    }
+
+    for branch in hg.branches():
+        branches[branch.name] = {
+            "target": node_id_2_swhid[branch.node_id],
+            "target_type": TargetType.REVISION.value,
+        }
+
+    for release in releases:
+        branches[release.name] = {
+            "target": release.swhid,
+            "target_type": TargetType.RELEASE.value,
+        }
+
+    return Snapshot.from_dict({"branches": branches}).id
+
+
+@main.command()
+@click.argument("rev", required=False)
+@click.pass_context
+def revision(ctx, rev):
+    """Compute the SWHID of a given revision.
+
+    If specified REV allow to select a single or multiple revisions
+    (using the Mercurial revsets language: `hg help revsets`)
+    """
+    hg = Hg(ctx.obj["HG_ROOT"])
+
+    for identity in identify_revision(hg, rev):
+        click.echo(identity)
+
+
+@main.command()
+@click.pass_context
+def snapshot(ctx):
+    """Compute the SWHID of the snapshot."""
+    root = ctx.obj["HG_ROOT"]
+    hg = Hg(root)
+
+    snapshot_swhid = identify_snapshot(hg)
+
+    uri = swhid("snapshot", snapshot_swhid)
+    click.echo(f"{uri}\t{root}")
+
+
+@main.command()
+@click.pass_context
+def all(ctx):
+    """Compute the SWHID of all the repository objects."""
+    root = ctx.obj["HG_ROOT"]
+    hg = Hg(root)
+
+    dir_uris = []
+    rev_uris = []
+    rel_uris = []
+
+    node_id_2_swhid = {}
+    for revision in identify_revision(hg):
+        dir_uris.append(revision.dir_uri())
+        rev_uris.append(str(revision))
+        node_id_2_swhid[revision.node_id] = revision.swhid
+
+    releases = []
+    for release in identify_release(hg, node_id_2_swhid):
+        rel_uris.append(str(release))
+        releases.append(release)
+
+    snapshot_swhid = identify_snapshot(hg, node_id_2_swhid, releases)
+
+    for uri in dir_uris + rev_uris + rel_uris:
+        click.echo(uri)
+
+    uri = swhid("snapshot", snapshot_swhid)
+    click.echo(f"{uri}\t{root}")
+
+
+if __name__ == "__main__":
+    main()
diff --git a/swh/loader/mercurial/tests/data/build.py b/swh/loader/mercurial/tests/data/build.py
new file mode 100755
--- /dev/null
+++ b/swh/loader/mercurial/tests/data/build.py
@@ -0,0 +1,187 @@
+#!/usr/bin/env python
+# Copyright (C) 2018-2020  The Software Heritage developers
+# See the AUTHORS file at the top-level directory of this distribution
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+import json
+import shutil
+import subprocess
+from datetime import datetime
+from pathlib import Path
+
+import click
+
+
+def abort(message):
+    """Abort the script with a message."""
+    click.echo(message, err=True)
+    click.get_current_context().abort()
+
+
+def backup(path: Path):
+    """Rename an existing path."""
+    click.echo(f"Creating backup of {path}")
+    now = datetime.now()
+    backup_path = path.with_suffix(f"{path.suffix}.bak.{now:%Y%m%d%H%M%S}")
+    path.rename(backup_path)
+    click.echo(f"Backup created: {str(backup_path)!r}")
+
+
+def _build_repository(script: str) -> Path:
+    """Build a repository from a bash script."""
+    script_path = Path(script).absolute()
+
+    if not script_path.exists():
+        abort(f"Path {script_path!r} does not exists.")
+
+    if script_path.suffix != ".sh":
+        abort(f"Wrong suffix: {script_path.suffix!r}. Expected: '.sh'")
+
+    repository_path = script_path.with_suffix("")
+
+    if repository_path.exists():
+        backup(repository_path)
+
+    click.echo(f"Running build script: {str(script_path)!r}")
+    subprocess.call(
+        ["bash", "-euo", "pipefail", script_path], env={"HG_REPO": str(repository_path)}
+    )
+
+    return repository_path
+
+
+def _build_json(source: str) -> Path:
+    if source.endswith(".tgz"):
+        archive_path = Path(source).absolute()
+        repository_path = archive_path.with_suffix("")
+
+        if repository_path.exists():
+            backup(repository_path)
+
+        subprocess.call(["tar", "-xf", archive_path], cwd=archive_path.parent)
+    elif source.endswith(".sh"):
+        repository_path = _build_repository(source)
+    else:
+        repository_path = Path(source).absolute()
+
+    click.echo(f"Extracting object identities: {str(repository_path)!r}")
+    output = subprocess.check_output(["swh-hg-identify", "all"], cwd=repository_path)
+    lines = output.decode().splitlines()
+
+    directory_swhids = []
+    revision_swhids = []
+    release_swhids = []
+
+    for line in lines:
+        uri, _ = line.split("\t")
+        _, _, swhid_type, swhid = uri.split(":")
+        if swhid_type == "dir":
+            directory_swhids.append(swhid)
+        elif swhid_type == "rev":
+            revision_swhids.append(swhid)
+        elif swhid_type == "rel":
+            release_swhids.append(swhid)
+        elif swhid_type == "snp":
+            snapshot_swhid = swhid
+        else:
+            abort(f"{line!r} unknown type {swhid_type!r}")
+
+    json_path = repository_path.with_suffix(".json")
+
+    if json_path.exists():
+        backup(json_path)
+
+    click.echo(f"Creating object identities file: {str(json_path)!r}")
+    json_path.write_text(
+        json.dumps(
+            {
+                "directories": directory_swhids,
+                "revisions": revision_swhids,
+                "releases": release_swhids,
+                "snapshot": snapshot_swhid,
+            }
+        )
+    )
+
+    return json_path
+
+
+@click.group()
+def main():
+    """Build example repositories archive from bash scripts."""
+
+
+@main.command("repository")
+@click.argument("script")
+def build_repository(script: str):
+    """Build a repository.
+
+    SCRIPT must be is a bash script with a `.sh` suffix
+
+    The generated repository will have the same path minor the `.sh` suffix.
+
+    The script will be executed in an already initialized repository.
+    So it only need to execute commands to populate the repository.
+    """
+    _build_repository(script)
+
+
+@main.command("json")
+@click.argument("source")
+def build_json(source: str):
+    """Build a json file of object identities.
+
+    SOURCE can be a script as required by the `repository` command
+    (see repository --help), a repository archive, or an existing repository.
+
+    The produced file will have the source path the `.json` suffix.
+    """
+    _build_json(source)
+
+
+@main.command("archive")
+@click.option(
+    "--clean", "-c", default=False, is_flag=True, help="Remove created artifacts",
+)
+@click.argument("source")
+def build_archive(source: str, clean: bool = False):
+    """Build a repository archive.
+
+    SOURCE can be a script as required by the `repository` command
+    (see repository --help), or an existing repository.
+
+    The produced archive will have the source path with the `.tgz` suffix.
+    It will contain the repository along with the json file of object identities.
+    """
+    if source.endswith(".sh"):
+        repository_path = _build_repository(source)
+    else:
+        repository_path = Path(source).absolute()
+        if not (repository_path / ".hg").exists():
+            abort(f"{str(repository_path)!r} is not a Mercurial repository")
+
+    json_path = _build_json(str(repository_path))
+
+    archive_path = repository_path.with_suffix(".tgz")
+    if archive_path.exists():
+        backup(archive_path)
+
+    subprocess.call(
+        [
+            "tar",
+            "-cf",
+            archive_path.relative_to(archive_path.parent),
+            repository_path.relative_to(archive_path.parent),
+            json_path.relative_to(archive_path.parent),
+        ],
+        cwd=archive_path.parent,
+    )
+
+    if clean:
+        shutil.rmtree(repository_path)
+        json_path.unlink()
+
+
+if __name__ == "__main__":
+    main()
diff --git a/swh/loader/mercurial/tests/data/hello.json b/swh/loader/mercurial/tests/data/hello.json
new file mode 100644
--- /dev/null
+++ b/swh/loader/mercurial/tests/data/hello.json
@@ -0,0 +1 @@
+{"directories": ["43d727f2f3f2f7cb3b098ddad1d7038464a4cee2", "8f2be433c945384c85920a8e60f2a68d2c0f20fb", "b3f85f210ff86d334575f64cb01c5bf49895b63e"], "revisions": ["8dd3db5d5519e4947f035d141581d304565372d2", "93b48d515580522a05f389bec93227fc8e43d940", "c3dbe4fbeaaa98dd961834e4007edb3efb0e2a27"], "releases": ["515c4d72e089404356d0f4b39d60f948b8999140"], "snapshot": "d35668e02e2ba4321dc951cd308cf883786f918a"}
\ No newline at end of file
diff --git a/swh/loader/mercurial/tests/data/the-sandbox.json b/swh/loader/mercurial/tests/data/the-sandbox.json
new file mode 100644
--- /dev/null
+++ b/swh/loader/mercurial/tests/data/the-sandbox.json
@@ -0,0 +1 @@
+{"directories": ["180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "180bd57623a7c2c47a8c43514a5f4d903503d0aa", "9cd8160c67ac4b0bc97e2e2cd918a580425167d3", "e2e117569b086ceabeeedee4acd95f35298d4553"], "revisions": ["17a62618eb6e91a1d5d8e1246ccedae020d3b222", "18012a93d5aadc331c468dac84b524430f4abc19", "1ee770fd10ea2d8c4f6e68a1dbe79378a86611e0", "24f45e41637240b7f9e16d2791b5eacb4a406d0f", "25f5b27dfa5ed15d336188ef46bef743d88327d4", "2652147529269778757d96e09aaf081695548218", "2973e5dc9568ac491b198f6b7f10c44ddc04e0a3", "2d4a801c9a9645fcd3a9f4c06418d8393206b1f3", "31cd7c5f669868651c57e3a2ba25ac45f76fa5cf", "32eb0354a660128e205bf7c3a84b46040ef70d92", "34192ceef239b8b72141efcc58b1d7f1676a18c9", "3565e7d385af0745ec208d719e469c2f58be8e94", "3ed4b85d30401fe32ae3b1d650f215a588293a9e", "40def747398c76ceec1bd248e3a6cb2a52e22dc5", "4d640e8064fe69b4c851dfd43915c431e80c7497", "4e2dc6d6073f0b6d348f84ded52f9143b10344b9", "4ef794980f820d44be94b2f0d53eb34d4241638c", "5017ce0b285351da09a2029ea2cf544f79b593c7", "553b09724bd30d9691b290e157b27a73e2d3e537", "5ee9ea92ed8cc1737b7670e39dab6081c64f2598", "5f4eba626c3f826820c4475d2d81410759ec911b", "61d762d65afb3150e2653d6735068241779c1fcf", "62ff4741eac1821190f6c2cdab7c8a9d7db64ad0", "6910964416438ca8d1698f6295871d727c4d4851", "70e750bb046101fdced06f428e73fee471509c56", "74335db9f45a5d1c8133ff7a7db5ed7a8d4a197b", "769db00b34b9e085dc699c8f1550c95793d0e904", "88b80615ed8561be74a700b92883ec0374ddacb0", "94be9abcf9558213ff301af0ecd8223451ce991d", "9c9e0ff08f215a5a5845ce3dbfc5b48c8050bdaf", "9e912851eb64e3a1e08fbb587de7a4c897ce5a0a", "9f82d95bd3edfb7f18b1a21d6171170395ea44ce", "a1f000fb8216838aa2a120738cc6c7fef2d1b4d8", "a41e2a548ba51ee47f22baad8e88994853d3e2f5", "a701d39a17a9f48c61a06eee08bd9ac0b8e3838b", "a9c4534552df370f43f0ef97146f393ef2f2a08c", "aafb69fd7496ca617f741d38c40808ff2382aabe", "b6932cb7f59e746899e4804f3d496126d1343615", "be34b8c7857a6c04e41cc06b26338d8e59cb2601", "be44d5e6cc66580f59c108f8bff5911ee91a22e4", "bec4c0a31b0b2502f44f34aeb9827cd090cca621", "c313df50bfcaa773dcbe038d00f8bd770ba997f8", "c346f6ff7f42f2a8ff867f92ab83a6721057d86c", "c4a95d5097519dedac437fddf0ef775136081241", "c77e776d22548d47a8d96463a3556172776cd59b", "c875bad563a73a25c5f3379828b161b1441a7c5d", "caef0cb155eb6c55215aa59aabe04a9c702bbe6a", "cb36b894129ca7910bb81c457c72d69d5ff111bc", "d2164061453ecb03d4347a05a77db83f706b8e15", "dafa445964230e808148db043c126063ea1dc9b6", "db9e625ba90056304897a94c92e5d27bc60f112d", "dc3e3ab7fe257d04769528e5e17ad9f1acb44659", "dcba06661c607fe55ec67b1712d153b69f65e38c", "dcddcc32740d2de0e1403e21a5c4ed837b352992", "ddecbc16f4c916c39eacfcb2302e15a9e70a231e", "e326a7bbb5bc00f1d8cacd6108869dedef15569c", "e874cd5967efb1f45282e9f5ce87cc68a898a6d0", "f2afbb94b319ef5d60823859875284afb95dcc18"], "releases": [], "snapshot": "3b8fe58e467deb7597b12a5fd3b2c096b8c02028"}
\ No newline at end of file
diff --git a/swh/loader/mercurial/tests/data/transplant.json b/swh/loader/mercurial/tests/data/transplant.json
new file mode 100644
--- /dev/null
+++ b/swh/loader/mercurial/tests/data/transplant.json
@@ -0,0 +1 @@
+{"directories": ["42a6ed8f073f00bb18114bc6228360283d26aef2", "615b36018bc3f8cc5a94ba34cb6fcd06b8a0cce7", "657ac477edb7ff761b33d1ebe53df29c12e21ca7", "6c3b565c034591f9e1ba9e7197ec6af62ed1ac8c", "96ee448816b927c395aa87a48734a41ab9a801b9", "c321c30480f216b818c32bbf7f0a5c728faa42cd"], "revisions": ["07589281b64120558940e2e38729b0decf16a88a", "2153ae1c0ac7825aa4fbf82647fa6548cb886546", "2e10f90a4e30ce3a07f1a11cc41e007b1def0bc1", "a95327fed1b5d6db6c7d5ad83621cb61f0f5f7d8", "c75e5af8f4b49ebc622d815459c88eda35ab050d", "ee83768fd9aadc306f835fc8f7caadc1cdc0e3df"], "releases": [], "snapshot": "42e9007138d3834723ffddaedea9139edb576036"}
\ No newline at end of file
diff --git a/swh/loader/mercurial/tests/dvcs_loader_checker.py b/swh/loader/mercurial/tests/dvcs_loader_checker.py
new file mode 100644
--- /dev/null
+++ b/swh/loader/mercurial/tests/dvcs_loader_checker.py
@@ -0,0 +1,132 @@
+import json
+from pathlib import Path
+from typing import List, NamedTuple, Set
+
+from swh.loader.core.loader import DVCSLoader
+from swh.model.model import Directory, Release, Revision, Snapshot
+
+
+class ExpectedSwhids(NamedTuple):
+    """List the of swhids expected from the loader."""
+
+    directories: Set[str]  # hex swhid of the root directory of each revision
+    revisions: Set[str]  # hex swhid of each revision
+    releases: Set[str]  # hex swhid of each release
+    snapshot: str  # hex swhid of the snapshot
+
+    @staticmethod
+    def load(path: Path) -> "ExpectedSwhids":
+        """Load expected swhids from a json file.
+
+        See `build.py` in the data directory on how to extract that json file
+        from an existing repository or archive.
+        """
+        data = json.load(open(path))
+        return ExpectedSwhids(
+            directories=set(data["directories"]),
+            revisions=set(data["revisions"]),
+            releases=set(data["releases"]),
+            snapshot=data["snapshot"],
+        )
+
+
+class DVCSLoaderChecker:
+    """Check the swhids produced by a DVCSLoader."""
+
+    def __init__(self, loader: DVCSLoader, expected: ExpectedSwhids) -> None:
+        self._loader = loader
+        self._expected = expected
+
+        self._directories: List[Directory] = []
+        self._revisions: List[Revision] = []
+        self._releases: List[Release] = []
+        self._snapshot: Snapshot
+
+        self._wrap_get_directories()
+        self._wrap_get_revisions()
+        self._wrap_get_releases()
+        self._wrap_get_snapshot()
+
+    def _wrap_get_directories(self):
+        """Replace the original loader's get_directories method to inspect its
+        output.
+        """
+        get_directories = self._loader.get_directories
+
+        def wrapper(*args, **kwargs):
+            for directory in get_directories(*args, **kwargs):
+                self._directories.append(directory)
+                yield directory
+            self._check_directories()
+
+        self._loader.get_directories = wrapper
+
+    def _check_directories(self):
+        """Check the emitted root directories.
+
+        Only check that the expected directories are present as only
+        root directories are expected.
+        """
+        swhids = set(directory.id.hex() for directory in self._directories)
+        assert (self._expected.directories - swhids) == set()
+
+    def _wrap_get_revisions(self):
+        """Replace the original loader's get_revisions method to inspect its
+        output.
+        """
+        get_revisions = self._loader.get_revisions
+
+        def wrapper(*args, **kwargs):
+            for revision in get_revisions(*args, **kwargs):
+                self._revisions.append(revision)
+                yield revision
+            self._check_revisions()
+
+        self._loader.get_revisions = wrapper
+
+    def _check_revisions(self):
+        """Check the emitted revisions."""
+        assert self._expected.revisions == set(
+            revision.id.hex() for revision in self._revisions
+        )
+
+    def _wrap_get_releases(self):
+        """Replace the original loader's get_releases method to inspect its
+        output.
+        """
+        get_releases = self._loader.get_releases
+
+        def wrapper(*args, **kwargs):
+            for release in get_releases(*args, **kwargs):
+                self._releases.append(release)
+                yield release
+            self._check_releases()
+
+        self._loader.get_releases = wrapper
+
+    def _check_releases(self):
+        """Check the emitted releases."""
+        assert self._expected.releases == set(
+            release.id.hex() for release in self._releases
+        )
+
+    def _wrap_get_snapshot(self):
+        """Replace the original loader's get_snapshot method to inspect its
+        output.
+        """
+        get_snapshot = self._loader.get_snapshot
+
+        def wrapper(*args, **kwargs):
+            self._snapshot = get_snapshot(*args, **kwargs)
+            self._check_snapshot()
+            return self._snapshot
+
+        self._loader.get_snapshot = wrapper
+
+    def _check_snapshot(self):
+        """Check the emitted snapshot."""
+        assert self._expected.snapshot == self._snapshot.id.hex()
+
+    def check(self):
+        """Check loader's outputs."""
+        assert self._loader.load() == {"status": "eventful"}
diff --git a/swh/loader/mercurial/tests/test_identify.py b/swh/loader/mercurial/tests/test_identify.py
new file mode 100644
--- /dev/null
+++ b/swh/loader/mercurial/tests/test_identify.py
@@ -0,0 +1,88 @@
+import os
+from urllib.parse import urlsplit
+
+from click.testing import CliRunner
+
+from swh.loader.mercurial.identify import main
+from swh.loader.tests import prepare_repository_from_archive
+
+
+def test_all_revisions(datadir: str, tmp_path: str):
+    archive_name = "hello"
+    archive_path = os.path.join(datadir, f"{archive_name}.tgz")
+    repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path)
+    directory = urlsplit(repo_url).path
+
+    runner = CliRunner()
+    result = runner.invoke(main, ["-d", directory, "revision"])
+
+    expected = (
+        "\n".join(
+            [
+                "swh:1:rev:93b48d515580522a05f389bec93227fc8e43d940"
+                "\t0a04b987be5ae354b710cefeba0e2d9de7ad41a9",
+                "swh:1:rev:8dd3db5d5519e4947f035d141581d304565372d2"
+                "\t82e55d328c8ca4ee16520036c0aaace03a5beb65",
+                "swh:1:rev:c3dbe4fbeaaa98dd961834e4007edb3efb0e2a27"
+                "\tb985ae4a07e12ac662f45a171e2d42b13be5b50c",
+            ]
+        )
+        + "\n"
+    )
+    assert result.output == expected
+
+
+def test_single_revision(datadir: str, tmp_path: str):
+    archive_name = "hello"
+    archive_path = os.path.join(datadir, f"{archive_name}.tgz")
+    repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path)
+    directory = urlsplit(repo_url).path
+
+    runner = CliRunner()
+    result = runner.invoke(
+        main, ["-d", directory, "revision", "0a04b987be5ae354b710cefeba0e2d9de7ad41a9"]
+    )
+
+    expected = (
+        "\n".join(
+            [
+                "swh:1:rev:93b48d515580522a05f389bec93227fc8e43d940"
+                "\t0a04b987be5ae354b710cefeba0e2d9de7ad41a9",
+            ]
+        )
+        + "\n"
+    )
+    assert result.output == expected
+
+
+def test_all(datadir: str, tmp_path: str):
+    archive_name = "hello"
+    archive_path = os.path.join(datadir, f"{archive_name}.tgz")
+    repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path)
+    directory = urlsplit(repo_url).path
+
+    runner = CliRunner()
+    result = runner.invoke(main, ["-d", directory, "all"])
+
+    expected = (
+        "\n".join(
+            [
+                "swh:1:dir:43d727f2f3f2f7cb3b098ddad1d7038464a4cee2"
+                "\t0a04b987be5ae354b710cefeba0e2d9de7ad41a9",
+                "swh:1:dir:b3f85f210ff86d334575f64cb01c5bf49895b63e"
+                "\t82e55d328c8ca4ee16520036c0aaace03a5beb65",
+                "swh:1:dir:8f2be433c945384c85920a8e60f2a68d2c0f20fb"
+                "\tb985ae4a07e12ac662f45a171e2d42b13be5b50c",
+                "swh:1:rev:93b48d515580522a05f389bec93227fc8e43d940"
+                "\t0a04b987be5ae354b710cefeba0e2d9de7ad41a9",
+                "swh:1:rev:8dd3db5d5519e4947f035d141581d304565372d2"
+                "\t82e55d328c8ca4ee16520036c0aaace03a5beb65",
+                "swh:1:rev:c3dbe4fbeaaa98dd961834e4007edb3efb0e2a27"
+                "\tb985ae4a07e12ac662f45a171e2d42b13be5b50c",
+                "swh:1:rel:515c4d72e089404356d0f4b39d60f948b8999140\t0.1",
+                f"swh:1:snp:d35668e02e2ba4321dc951cd308cf883786f918a\t{directory}",
+            ]
+        )
+        + "\n"
+    )
+    assert result.output == expected
diff --git a/swh/loader/mercurial/tests/test_loader.py b/swh/loader/mercurial/tests/test_loader.py
--- a/swh/loader/mercurial/tests/test_loader.py
+++ b/swh/loader/mercurial/tests/test_loader.py
@@ -22,6 +22,18 @@
 from swh.storage.algos.snapshot import snapshot_get_latest
 
 from ..loader import CloneTimeoutError, HgArchiveBundle20Loader, HgBundle20Loader
+from .dvcs_loader_checker import DVCSLoaderChecker, ExpectedSwhids
+
+
+def test_examples(swh_config, datadir, tmp_path):
+    for archive_name in ("hello", "transplant", "the-sandbox"):
+        archive_path = os.path.join(datadir, f"{archive_name}.tgz")
+        json_path = os.path.join(datadir, f"{archive_name}.json")
+        repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path)
+
+        DVCSLoaderChecker(
+            loader=HgBundle20Loader(repo_url), expected=ExpectedSwhids.load(json_path),
+        ).check()
 
 
 def test_loader_hg_new_visit_no_release(swh_config, datadir, tmp_path):