Request clarification for qdrant imports and update for README page

[EdNett]

We wish to use Qdrant 1.16.2 and in mode gRPC - and we don't know how to import properly, because there is no clear explanation of how to do that, and it appears it is constantly changing. For example we have -

from qdrant_client import QdrantClient
from qdrant_client.http.models import (
Distance,
VectorParams,
Range,
PointStruct,
PointIdsList,
Filter,
FilterSelector,
FieldCondition,
MatchValue,
ScoredPoint,
Payload,
)

from qdrant_client.conversions.common_types import PointId

currently because PointID was throwing errors when we tried to import from qdrant_client.http.models .

On your README page you suggest : qdrant_client import ... and qdrant_client.models import ...

Formnerly we were running in http mode (gRPC=False) on port 6333 only. Now since we wanted gRPC, and changed the port number and set gRPC=True :
When I restart our qdrant server I see that it seems to be running in both modes and I see this:

Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.095842Z INFO qdrant: Distributed mode disabled
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.095866Z INFO qdrant: Telemetry reporting enabled, id: 9f49c886-d78c-47c8-972e-08cfef1cdd75
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.103942Z WARN qdrant::actix::web_ui: Static content folder for Web UI './static' does not exist
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.104284Z INFO qdrant::actix: TLS disabled for REST API
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.104384Z INFO qdrant::actix: Qdrant HTTP listening on 6333
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.104426Z INFO actix_server::builder: starting 31 workers
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.104440Z INFO actix_server::server: Actix runtime found; starting in Actix runtime
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.104452Z INFO actix_server::server: starting service: "actix-web-service-0.0.0.0:6333", workers: 3>
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.116228Z INFO qdrant::tonic: Qdrant gRPC listening on 6334
Jan 11 10:38:17 ns3183956 qdrant[1540289]: 2026-01-11T10:38:17.116283Z INFO qdrant::tonic: TLS disabled for gRPC API

So, where can I find a clear guide on what to import from ? i.e. qdrant_client.http.models, qdrant_client.models, qdrant_client or qdrant_client.conversions.common_types or what?

Yes I am new here, but this could really use some clarity for all to use. This is with python3.14.2 . Is qdrant 1.16.2 compatible with python 3.14 ?

Thank you.

[joein]

Hey @EdNett

You can import from qdrant_client.models

You don't need to import conversions

Could you provide a code snippet of how you're trying to use grpc, so I could help you out better?

[EdNett]

Sure, thanks for any help> We had this error trying to import PointId : Jan 11 13:33:08 ns3183956 python3.14[1615215]: from qdrant_client.models import (
Jan 11 13:33:08 ns3183956 python3.14[1615215]: ...<12 lines>...
Jan 11 13:33:08 ns3183956 python3.14[1615215]: )
Jan 11 13:33:08 ns3183956 python3.14[1615215]: ImportError: cannot import name 'PointId' from 'qdrant_client.models' (/srv/custom/rag-web3/venv/lib/python3.14/site-packag>

Is PointId deprecated? Here is how I am using it:

def index_web_page(
url,
lang,
raw_html,
text,
chunks,
last_modified,
file_size,
content_type="text",
source_type="web",
):
embs = model.encode(chunks, batch_size=16, convert_to_numpy=True, normalize_embeddings=True)

points = []
for i, ch in enumerate(chunks):
    points.append(
        {
            "id": PointId(uuid=str(uuid.uuid4())),
            "vector": embs[i].tolist(),
            "payload": {
                "lang": lang,
                "url": url,
                "type": content_type,
                "text": ch,
                "source_path": url,
                "source_type": source_type,
                "last_modified": last_modified,
                "file_size": file_size,
            },
        }
    )

ensure_collection()
qdrant_client.upsert(collection_name=COLLECTION_NAME, points=points)
print(f"Indexed {len(points)} chunks from {url}")

Is there a newer way to do this without PointId ? I am getting these types of errors:

gunicorn[1649464]: [ERROR] /srv/custom/rag-web3/docs/General-legal-2023.md: 'typing.Union' object is not callable

#1140
Thanks for any help understanding whether or not PointID is deprecated.

[joein]

If you want to import a grpc model, then you need to import it from qdrant_client.grpc
However, you usually don't need to do that, just use http models and we do the conversion automatically

E.g.

from qdrant_client import models
point = models.PointStruct(id=uuid.uuid4(), vector=vector, payload=payload)

[uahroberto]

Hi @EdNett,

I took a look at this to reproduce the issue locally.

As @joein mentioned, PointId is not exposed via qdrant_client.models, which causes the ImportError you are seeing. It seems intended for internal gRPC conversions rather than direct usage.

To fix both the import error and the typing.Union error you mentioned later, the best approach is to wrap your data in models.PointStruct. This ensures everything is correctly typed before sending it to the client.

Here is how you can adapt your loop:

from qdrant_client import models
import uuid

# ... inside your indexing function ...

points = []
for i, ch in enumerate(chunks):
    points.append(
        models.PointStruct(
            id=str(uuid.uuid4()), 
            vector=embs[i].tolist(),
            payload={
                "lang": lang,
                "url": url,
                "type": content_type,
                "text": ch,
                "source_path": url,
                "source_type": source_type,
                "last_modified": last_modified,
                "file_size": file_size,
            }
        )
    )

# Now upsert will work without issues
qdrant_client.upsert(collection_name=COLLECTION_NAME, points=points)

I tested this locally with the latest client version and it works as expected. Hope this helps!