SAM3

Use Meta's SAM3 model through our Serverless Cloud API

We support Meta's Segment Anything Model 3 inferencing via our Serverless Cloud API. We offer two different SAM3 endpoints:

Training a SAM3 model on Roboflow is available on paid plans that include usage-based billing. From there, you can request access with the "Request Feature" button on the SAM3 architecture to use the feature training flow.

Fine-tuned SAM3 models cannot run on the Serverless Cloud API. Deploy them on a Dedicated Deployment or self-hosted Inference. The hosted sam3 endpoints on this page are unaffected.

• Promptable concept segmentation (PCS), which segments every instance of a concept in the image. Concepts are described by text prompts, exemplar boxes, or both.

• Promptable visual segmentation (PVS), which interactively segments one object per request from points or a box, in the style of SAM2.

Use this table to pick an endpoint:

You have	You want	Use
A text description (ex: "person")	Masks for every matching instance	/sam3/concept_segment
A box around one example object	Masks for every similar instance	/sam3/concept_segment
Text plus example boxes to include or exclude objects	Masks for every matching instance	/sam3/concept_segment
A click or a box on one specific object	A mask for that object only	/sam3/visual_segment

Pass your API key as the api_key query parameter on every request.

Concept Segmentation (PCS)

POST https://serverless.roboflow.com/sam3/concept_segment

Each entry in prompts describes one concept. The response contains one prompt_results entry per prompt, each holding every instance found. Requests accept at most 16 prompts.

Text prompts

import os
import requests

payload = {
    "image": {"type": "url", "value": "https://media.roboflow.com/inference/people-walking.jpg"},
    "prompts": [
        {"type": "text", "text": "person"},
        {"type": "text", "text": "backpack"},
    ],
    "output_prob_thresh": 0.5,
    "format": "polygon",  # or "rle"
}

response = requests.post(
    "https://serverless.roboflow.com/sam3/concept_segment",
    params={"api_key": os.environ["ROBOFLOW_API_KEY"]},
    json=payload,
)
for prompt_result in response.json()["prompt_results"]:
    print(prompt_result["echo"], len(prompt_result["predictions"]), "instances")

Images can also be sent inline as {"type": "base64", "value": "<BASE64_IMAGE>"}.

Exemplar box prompts

Instead of text, you can prompt with an exemplar: a box around one example object. The model finds every instance that matches the example, not just the boxed object.

payload = {
    "image": {"type": "url", "value": "https://media.roboflow.com/inference/people-walking.jpg"},
    "prompts": [
        {
            "type": "visual",
            "boxes": [{"x": 1409, "y": 705, "width": 112, "height": 183}],
            "box_labels": [1],
        }
    ],
    "output_prob_thresh": 0.5,
    "format": "polygon",
}

Boxes use absolute pixel coordinates. Two formats are accepted:

• {"x": ..., "y": ..., "width": ..., "height": ...} where x, y is the top-left corner

• {"x0": ..., "y0": ..., "x1": ..., "y1": ...} for explicit corners

box_labels is required when boxes is set and must have one entry per box: 1 marks a positive exemplar (find objects like this), 0 marks a negative exemplar (exclude objects like this).

Combined text and exemplar prompts

A single prompt can carry both text and exemplar boxes. This is useful for narrowing a text concept with visual examples, or excluding lookalikes with negative exemplars:

payload = {
    "image": {"type": "url", "value": "https://media.roboflow.com/inference/people-walking.jpg"},
    "prompts": [
        {
            "type": "visual",
            "text": "person",
            "boxes": [
                {"x": 1409, "y": 705, "width": 112, "height": 183},
                {"x": 1216, "y": 496, "width": 124, "height": 184},
            ],
            "box_labels": [1, 0],
        }
    ],
    "output_prob_thresh": 0.5,
    "format": "polygon",
}

Here the model segments people matching the first (positive) exemplar while suppressing instances similar to the second (negative) exemplar.

Visual Segmentation (PVS)

POST https://serverless.roboflow.com/sam3/visual_segment

PVS segments one specific object indicated by clicks or a box. Use it for interactive, human-in-the-loop mask refinement; use PCS when you want every instance of a concept.

import os
import requests

payload = {
    "image": {"type": "url", "value": "https://media.roboflow.com/inference/people-walking.jpg"},
    "prompts": {
        "prompts": [
            {
                "points": [{"x": 1465, "y": 796, "positive": True}],
                "box": {"x": 1465, "y": 796, "width": 112, "height": 183},
            }
        ]
    },
    "multimask_output": False,
    "format": "json",
}

response = requests.post(
    "https://serverless.roboflow.com/sam3/visual_segment",
    params={"api_key": os.environ["ROBOFLOW_API_KEY"]},
    json=payload,
)
prediction = response.json()["predictions"][0]
print(prediction["confidence"], len(prediction["masks"]), "polygons")

A prompt can contain points, a box, or both:

• points are absolute pixel coordinates. "positive": true includes the clicked region, false excludes it. Add more points to refine the mask.

• box uses center-anchored coordinates: x, y is the box center, unlike PCS boxes which are top-left anchored.

The response contains the single highest-confidence mask for the prompt. multimask_output controls how many internal mask proposals the model generates (three when true), but the best proposal is always selected for the response.

Send one prompt per request. Multiple prompts in one PVS request currently return only one prediction.

For an interactive demo using OpenCV, see this GitHub Gist, which was used in this video:

Inference speed

Latency measured with Roboflow Inference on 1x NVIDIA L4, batch size 1, mean after warmup.

Model	Latency (ms)
sam3	251.4

Measured with concept segmentation from a single text prompt.

Endpoints

SAM3 PCS (promptable concept segmentation)

post

Concept Segmentation (Text Prompts)

Allows you to segment objects using text prompts.

Image Input: The image field accepts either:

• {"type": "url", "value": "<IMAGE_URL>"} - A publicly accessible image URL

• {"type": "base64", "value": "<BASE64_DATA>"} - Base64 encoded image data

Prompts: Each prompt in the prompts array should have type: "text" and a text field with the object description.

Query parameters

api_keystringRequired

Your Roboflow API Key. Get one at https://app.roboflow.com/settings/api

Body

formatstringOptional

One of 'polygon', 'rle'

Default: polygon

image_idstringOptional

Optional ID for caching embeddings.

output_prob_threshnumberOptional

Score threshold for outputs.

Default: 0.5

model_idstringOptional

The model ID of SAM3. Use 'sam3/sam3_final' to target the generic base model.

Default: sam3/sam3_final

nms_iou_thresholdnumberOptional

IoU threshold for cross-prompt NMS. If not set, NMS is disabled. Must be in [0.0, 1.0] when set.

Responses

200

Successful Response

application/json

timenumberRequired

The time in seconds it took to produce the segmentation including preprocessing

422

Validation Error

application/json

post/sam3/concept_segment

POST /sam3/concept_segment?api_key=text HTTP/1.1
Host: serverless.roboflow.com
Content-Type: application/json
Accept: */*
Content-Length: 206

{
  "image": {
    "type": "url",
    "value": "https://media.roboflow.com/notebooks/examples/dog.jpeg"
  },
  "prompts": [
    {
      "type": "text",
      "text": "person"
    },
    {
      "type": "text",
      "text": "car"
    }
  ],
  "output_prob_thresh": 0.5,
  "format": "polygon"
}

{
  "prompt_results": [
    {
      "prompt_index": 0,
      "echo": {
        "prompt_index": 0,
        "type": "text",
        "text": "dog",
        "num_boxes": 0
      },
      "predictions": [
        {
          "masks": [
            [
              [
                345,
                251
              ],
              [
                344,
                252
              ],
              [
                343,
                253
              ]
            ]
          ],
          "confidence": 0.89453125,
          "format": "polygon"
        }
      ]
    }
  ],
  "time": 0.221
}

SAM3 PVS (promptable visual segmentation)

post

Interactive Segmentation (SAM 2 Style)

SAM 3 also supports interactive segmentation using points and boxes.

Image Input: The image field accepts either:

• {"type": "url", "value": "<IMAGE_URL>"} - A publicly accessible image URL

• {"type": "base64", "value": "<BASE64_DATA>"} - Base64 encoded image data

Note: NumPy arrays are NOT supported on the serverless API. Use URL or base64 encoding only.

Prompts: Support point-based prompts with positive/negative clicks for interactive segmentation.

Query parameters

api_keystringRequired

Your Roboflow API Key. Get one at https://app.roboflow.com/settings/api

Body

SAM2 visual segmentation request.

image_idstringOptional

The ID of the image to be segmented used to retrieve cached embeddings. If an embedding is cached, it will be used instead of generating a new embedding. If no embedding is cached, a new embedding will be generated and cached.

Example: image_id

formatstringOptional

The format of the response. Must be one of 'json', 'rle', or 'binary'. If binary, masks are returned as binary numpy arrays. If json, masks are converted to polygons. If rle, masks are converted to RLE format.

Default: jsonExample: json

sam2_version_idstringOptional

The version ID of SAM to be used for this request. Must be one of hiera_tiny, hiera_small, hiera_large, hiera_b_plus

Default: hiera_largeExample: hiera_large

multimask_outputbooleanOptional

If true, the model will return three masks. For ambiguous input prompts (such as a single click), this will often produce better masks than a single prediction.

Default: trueExample: true

save_logits_to_cachebooleanOptional

If True, saves the low-resolution logits to the cache for potential future use.

Default: false

load_logits_from_cachebooleanOptional

If True, attempts to load previously cached low-resolution logits for the given image and prompt set.

Default: false

Responses

200

Successful Response

application/json

timenumberRequired

The time in seconds it took to produce the segmentation including preprocessing

422

Validation Error

application/json

post/sam3/visual_segment

POST /sam3/visual_segment?api_key=text HTTP/1.1
Host: serverless.roboflow.com
Content-Type: application/json
Accept: */*
Content-Length: 294

{
  "image": {
    "type": "url",
    "value": "http://www.example-image-url.com"
  },
  "image_id": "image_id",
  "prompts": [
    {
      "prompts": [
        {
          "points": [
            {
              "positive": true,
              "x": 100,
              "y": 100
            }
          ]
        }
      ]
    }
  ],
  "format": "json",
  "sam2_version_id": "hiera_large",
  "multimask_output": true,
  "save_logits_to_cache": false,
  "load_logits_from_cache": false
}

{
  "prompt_results": [
    {
      "prompt_index": 1,
      "predictions": []
    }
  ],
  "time": 1
}

Use with Inference (self-hosted)

SAM3 can also run on your own hardware, either loaded in-process with the inference package or served from a GPU container.

Run in Docker

docker run -it --rm -p 9001:9001 --gpus=all roboflow/inference-server:latest

The server exposes the same /sam3/concept_segment and /sam3/visual_segment endpoints documented above at http://localhost:9001.

Load the model in Python

pip install "inference-gpu[sam3]"

import os

os.environ["API_KEY"] = "YOUR_API_KEY"

from inference.core.entities.requests.sam3 import Sam3Prompt
from inference.models.sam3 import SegmentAnything3

model = SegmentAnything3(model_id="sam3/sam3_final")

prompts = [
    # Segment every instance of a concept
    Sam3Prompt(type="text", text="person"),
    # Box one example object and segment every similar instance.
    # box_labels: 1 = positive exemplar, 0 = negative exemplar.
    Sam3Prompt(
        type="visual",
        boxes=[Sam3Prompt.Box(x=1409, y=705, width=112, height=183)],
        box_labels=[1],
    ),
]

response = model.segment_image(
    image="path/to/your/image.jpg",
    prompts=prompts,
    output_prob_thresh=0.5,
    format="polygon",  # or "rle", "json"
)

for prompt_result in response.prompt_results:
    print(prompt_result.echo.text, len(prompt_result.predictions), "instances")

Weights download automatically on first use.

Interactive segmentation in Python

Sam3ForInteractiveImageSegmentation implements the SAM2-style point and box interface, for human-in-the-loop mask refinement:

from inference.models.sam3 import Sam3ForInteractiveImageSegmentation

model = Sam3ForInteractiveImageSegmentation(model_id="sam3/sam3_final")

embedding, img_shape, image_id = model.embed_image(image="path/to/image.jpg")

masks, scores, logits = model.segment_image(
    image_id=image_id,
    prompts={"points": [{"x": 500, "y": 400, "positive": True}]},
)

Use in Workflows

Two SAM3 image blocks are available in Workflows:

• SAM 3 runs concept segmentation. Enter the classes you want in class_names (for example ["person", "vehicle"]) and the block outputs instance segmentation predictions that other steps can consume.

• SAM 3 Interactive runs promptable visual segmentation. Supply labeled points (kind labeled_points), for example [{"x": 320, "y": 240, "positive": true}], and optionally connect detections from another model to the boxes field. Each box becomes a separate prompt, and its class name is forwarded to the predicted mask.

Video tracking

The SAM3 Video Tracker block (roboflow_core/sam3_video@v1) runs SAM3's streaming concept tracker frame by frame. You provide concepts as text in class_names, and the model runs fused detection and tracking on every frame. Objects matching a concept keep a stable tracker_id, and, unlike detector-seeded tracking, objects that enter the scene mid-stream are picked up automatically with no re-prompting and no upstream detection model. Each mask carries the concept it matched as its class name and the model's detection score as its confidence (filter with threshold, default 0.5).

• Stateful and local-only. One tracking session is kept per video_metadata.video_identifier. The block requires WORKFLOWS_STEP_EXECUTION_MODE=local, a GPU, and a persistent WebRTC session.

• No prompt scheduling. Concept prompts are registered once per session; the session is re-seeded only when the stream restarts or class_names changes. For detector-driven (box-prompted) video tracking, use the SAM2 Video Tracker block on the SAM2 page, which also accepts sam3trackervideo as model_id.

• Model. model_id defaults to sam3video, the HuggingFace transformers port of SAM3 video, which exposes the frame-by-frame streaming interface. The native sam3 package's video predictor requires the whole video upfront and cannot be used for live streams.

from inference_sdk import InferenceHTTPClient
from inference_sdk.webrtc import StreamConfig, VideoFileSource

WORKFLOW = {
    "version": "1.0",
    "inputs": [{"type": "InferenceImage", "name": "image"}],
    "steps": [
        {
            "type": "roboflow_core/sam3_video@v1",
            "name": "tracker",
            "images": "$inputs.image",
            "class_names": ["person", "forklift"],
            "threshold": 0.5,
        },
    ],
    "outputs": [
        {
            "type": "JsonField",
            "name": "predictions",
            "selector": "$steps.tracker.predictions",
        }
    ],
}

client = InferenceHTTPClient(
    api_url="http://localhost:9001",
    api_key="YOUR_API_KEY",
)

session = client.webrtc.stream(
    source=VideoFileSource("path/to/video.mp4"),
    workflow=WORKFLOW,
    config=StreamConfig(data_output=["predictions"]),
)

@session.on_data("predictions")
def handle_predictions(predictions, metadata):
    print(predictions)

session.run()

SAM3-3D (beta)

SAM3-3D turns a 2D image plus masks into 3D assets: meshes and Gaussian splats.

SAM3-3D is in beta. It is available only when the SAM3_3D_OBJECTS_ENABLED flag is set, requires a GPU with 32 GB or more of VRAM, and runs through the inference package or a local Inference server (it is not on the Serverless Cloud API).

Install the dependencies (Python 3.10 recommended):

pip install --no-cache-dir --no-build-isolation -r requirements/requirements.sam3_3d.txt

Or build and run the 3D-enabled GPU container:

docker build -t roboflow/roboflow-inference-server-gpu:dev -f docker/dockerfiles/Dockerfile.onnx.gpu.3d .
docker run --gpus all -p 9001:9001 roboflow/roboflow-inference-server-gpu:dev

Input. An RGB image plus mask_input, which defines the object regions. Masks are accepted as binary arrays ((H, W) or (N, H, W)), COCO flat polygons, point-pair polygons, RLE dicts, or an sv.Detections object from SAM2 or another segmentation model.

Output. mesh_glb (combined scene mesh, GLB), gaussian_ply (combined Gaussian splat, PLY), objects (per-object mesh_glb, gaussian_ply, and metadata with rotation, translation, and scale), and time.

import os

os.environ["SAM3_3D_OBJECTS_ENABLED"] = "true"
os.environ["SPARSE_ATTN_BACKEND"] = "flash_attn"
os.environ["ATTN_BACKEND"] = "flash_attn"

from inference import get_model
from inference.core.entities.requests.sam3_3d import Sam3_3D_Objects_InferenceRequest

model = get_model("sam3-3d-objects", api_key="YOUR_API_KEY")

request = Sam3_3D_Objects_InferenceRequest(
    image={"type": "file", "value": "image.jpg"},
    mask_input=mask_polygons,  # polygons, binary masks, or RLE
)

response = model.infer_from_request(request)

if response.mesh_glb is not None:
    with open("out_mesh.glb", "wb") as f:
        f.write(response.mesh_glb)

for index, obj in enumerate(response.objects):
    if obj.gaussian_ply is not None:
        with open(f"out_object_{index}.ply", "wb") as f:
            f.write(obj.gaussian_ply)

Setting SPARSE_ATTN_BACKEND and ATTN_BACKEND to flash_attn speeds up the pipeline. In Workflows, SAM3-3D supports local execution and remote execution through the sam3_3d_infer() client method or the /sam3_3d/infer endpoint.

See also

• SAM2 - point and box prompted segmentation, plus detector-seeded video tracking.

• Segment Anything (SAM) - the original single-object model.