roboflow-data-management

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese
For agents: This skill is authored in
roboflow/computer-vision-skills
. If your client already has it loaded as a skill (for example from a Roboflow plugin), use that copy. The
skill://roboflow/<name>/SKILL.md
MCP resources serve the same skills for clients that don't have them loaded.
适用于 Agent: 此 skill 由
roboflow/computer-vision-skills
编写。如果你的客户端已将其作为 skill 加载(例如通过 Roboflow plugin),请使用该副本。
skill://roboflow/<name>/SKILL.md
MCP resources 可为尚未加载这些 skills 的客户端提供相同内容。

Data Management on Roboflow

Roboflow 数据管理

Project Types

项目类型

TypeAnnotation FormatUse Case
Object DetectionBounding box (polygon/mask auto-converted)Locate objects with boxes
Instance SegmentationPolygon, MaskPixel-level per-object boundaries
Semantic SegmentationPolygon, MaskPixel-level class regions
Keypoint DetectionKeypoints (skeleton)Pose/skeleton estimation
Single-Label ClassificationImage-level label (no drawn annotations)One class per image
Multi-Label ClassificationImage-level labelsMultiple classes per image
Project type is set at creation and cannot be changed later.
类型标注格式使用场景
目标检测(Object Detection)Bounding box(polygon/mask 会自动转换)使用边界框定位对象
实例分割(Instance Segmentation)Polygon、Mask像素级的单对象边界
语义分割(Semantic Segmentation)Polygon、Mask像素级类别区域
关键点检测(Keypoint Detection)Keypoints(skeleton)姿态/骨架估计
单标签分类(Single-Label Classification)图像级标签(不绘制标注)每张图像一个类别
多标签分类(Multi-Label Classification)图像级标签每张图像多个类别
项目类型在创建时设定,之后无法更改。

Uploading Data

上传数据

Methods

方法

MethodBest ForFormats
Web UI drag-and-drop< 1,000 imagesJPG, PNG, WEBP, AVIF, BMP, MOV, MP4, PDF + 40+ annotation formats
CLI (
roboflow import
)
> 1,000 images (images only)Same image formats, no video
MCP host-assisted upload (
image_upload_files
)
Preferred when the host supplies the selected files in the tool's
files
input (ChatGPT attachments, Codex local paths), up to 100 files / 1 GB per call
Images plus annotations, MP4/MOV for video-only projects, or one dataset ZIP
MCP signed-URL upload (
image_upload
)
Fallback when you would have to invent
download_url
or
file_id
, or the upload exceeds 1 GB, and the client can ZIP and PUT files, up to 2 GB / 10,000 files per ZIP
Images plus annotations, or MP4/MOV for video-only projects
Cloud storage bucket mirrorExtensive or continuously-growing data already in S3 / GCSSee
skill://roboflow/roboflow-cloud-storage/SKILL.md
Dataset Upload Workflow BlockCollecting from production WorkflowsProgrammatic
Universe forkStarting from a public datasetAny Universe dataset
Limits: Max 20 MB per image, max 16,400 x 10,900 px. Duplicate images are skipped automatically.
方法最适用场景格式
Web UI 拖放少于 1,000 张图像JPG、PNG、WEBP、AVIF、BMP、MOV、MP4、PDF + 40 多种标注格式
CLI(
roboflow import
)
超过 1,000 张图像(仅限图像)相同的图像格式,不支持视频
MCP 宿主辅助上传(
image_upload_files
)
当宿主通过工具的
files
输入提供所选文件时应优先使用(ChatGPT 附件、Codex 本地路径),每次调用最多 100 个文件 / 1 GB
图像及标注文件;仅视频项目支持 MP4/MOV;或单个数据集 ZIP
MCP signed-URL 上传(
image_upload
)
当不得不自行构造
download_url
或
file_id
,或上传超过 1 GB,且客户端可以对文件进行 ZIP 打包并执行 PUT 时作为回退方案;每个 ZIP 最多 2 GB / 10,000 个文件
图像及标注文件;仅视频项目支持 MP4/MOV
云存储 bucket 镜像数据量较大或持续增长,且已存储在 S3 / GCS 中参见
skill://roboflow/roboflow-cloud-storage/SKILL.md
Dataset Upload Workflow Block从生产环境 Workflows 收集数据程序化
Universe fork从公共数据集开始任意 Universe 数据集
限制: 每张图像最大 20 MB,最大 16,400 x 10,900 px。重复图像会自动跳过。

Video Upload

视频上传

For native MP4/MOV uploads into a video-only project such as Action Recognition, prefer
image_upload_files
when the host supplies the videos in its
files
input. It packages the selected videos into a ZIP server-side; the platform stages each one as a native video. It does not extract frames. Image projects skip zipped videos, and video-only projects skip images; inspect the upload warnings.
When you would have to invent
download_url
or
file_id
, or host transfer cannot handle the selected video, and the client can PUT to storage, use
video_upload
for one raw MP4/MOV up to 3 GB, or
image_upload
for a video ZIP up to 2 GB. With
video_upload
, send every
requiredHeaders
entry and the video's Content-Type, then poll
video_upload_status
until
uploaded
or
failed
. Use the final
videoId
from the uploaded response; deduplication can change it.
For frames in an image project, use the web uploader's frame extraction or extract frames locally and upload the images. For Asset Library storage or media analysis without a project,
media_upload
remains a separate signed PUT +
media_upload_finalize
flow; it does not accept
fileParams
.
对于向仅视频项目(如 Action Recognition)原生上传 MP4/MOV,如果宿主在其
files
输入中提供视频,优先使用
image_upload_files
。它会在服务器端将所选视频打包为 ZIP;平台会将每个视频暂存为原生视频。它不会提取帧。图像项目会跳过 ZIP 中的视频,仅视频项目会跳过图像;请检查上传警告。
当不得不自行构造
download_url
或
file_id
,或宿主传输无法处理所选视频,并且客户端可以向存储执行 PUT 时,使用
video_upload
上传单个最大 3 GB 的原始 MP4/MOV,或使用
image_upload
上传最大 2 GB 的视频 ZIP。使用
video_upload
时,请发送
requiredHeaders
中的每一项以及视频的 Content-Type,然后轮询
video_upload_status
,直至状态为
uploaded
或
failed
。使用上传响应中的最终
videoId
;去重可能会改变该 ID。
对于图像项目中的帧,请使用 Web 上传器的帧提取功能,或在本地提取帧后上传图像。如果只需为 Asset Library 存储或进行媒体分析而不关联项目,
media_upload
仍采用独立的 signed PUT +
media_upload_finalize
流程;它不接受
fileParams
。

CLI Upload

CLI 上传

bash
pip install roboflow
roboflow import -w <workspace> -p <project-id> /path/to/dataset
bash
pip install roboflow
roboflow import -w <workspace> -p <project-id> /path/to/dataset

MCP Image and Video Upload

MCP 图像与视频上传

Use
image_upload_files
when the host supplies the user's selected files in the tool's
files
input. ChatGPT supplies attachment references, and Codex may accept absolute local paths and transfer them itself. If you would have to invent
download_url
or
file_id
, the host has no transfer bridge, so use the signed-URL flow below instead. Do not invent file references or pass file bytes through the model.
The Roboflow server downloads the files, assembles a ZIP, and uploads it to storage. The client does not need to ZIP files or PUT to GCS. Accepts images plus annotation files, MP4/MOV for video-only projects, or one dataset ZIP on its own, up to 100 files / 1 GB total per call. For larger file counts, use one ZIP within the host and tool limits. This is a project media import, not general document storage.
Use
image_upload
when the host does not supply the files in the
files
input or cannot handle them, and the client can ZIP and HTTP PUT to the returned signed URL. An authentication, validation, or server error alone does not establish that the bridge is unavailable. If the sandbox cannot reach storage, do not repeatedly retry the PUT; use host-assisted transfer or the project's web uploader.
After either flow returns a
taskId
(and after the client PUT for the fallback), poll
image_upload_status
until it is
completed
or
failed
. Inspect
result.uploadErrors
,
annotationErrors
and
warnings
, even when the task completed. For every
result.videoIds
entry, also poll
video_upload_status
until
uploaded
or
failed
;
videosStaged
and ZIP completion do not prove video ingestion succeeded. Poll any returned video IDs even if another ZIP chunk failed. Use each final uploaded
videoId
, not the pre-deduplication ID.
result.videoIds
stops at 100 IDs, and
result.videosStaged
only counts videos from chunks that succeeded. If
result.videosStaged
exceeds the returned IDs or
result.failedChunks
is above 0, the task did not return every video ID: poll the IDs it did return and report the rest as unverified, with the counts.
images_search(project_id, media_types=["video"])
searches the whole project, not this import, so it can return older videos and miss pending or failed ones. Use a search result only as supporting evidence when it matches an uploaded file by name, never as proof that the import is complete.
Both flows use the same image and annotation import behavior:
  • Annotations are imported. The zip can hold annotation files next to the images. They are parsed like in the web uploader: COCO JSON (
    _annotations.coco.json
    ), YOLO TXT (with
    data.yaml
    or
    classes.txt
    ), Pascal VOC XML, CreateML JSON, CSV, and more. Annotations are matched to images by file name. In classification projects, an image's parent folder name is its class.
  • Folders set splits. A
    train/
    ,
    valid/
    or
    test/
    folder (also
    training/
    ,
    validation/
    ,
    val/
    ,
    testing/
    ) anywhere in an image or video's path sets its split. Other files use the
    split
    argument.
  • Annotated images go straight into the Dataset. There is no annotation job to accept, unless the workspace has annotation review turned on (then they wait in a review job). Images without annotations land in a batch (
    batch_name
    , default "Uploaded via API"); label them with
    annotations_save
    and
    add_to_dataset=true
    .
  • Duplicates still get their annotations. An image already in the project is not stored again, but its annotation from the zip is saved. If it already has an annotation, that one is kept and the image is listed in
    annotationErrors
    as "Image was already annotated"; pass
    annotation_overwrite=true
    to replace it. Classification labels are always replaced.
  • A failed chunk is safe to retry. The zip is processed in chunks. If the status is
    failed
    with an
    uploadErrors
    entry like
    chunk-0
    , that chunk's images may be stored without their annotations. Upload the same zip again: stored images come back as duplicates and their missing annotations are saved. "Image was already annotated" for images that made it the first time is expected.
当宿主通过工具的
files
输入提供用户所选文件时,使用
image_upload_files
。ChatGPT 会提供附件引用,Codex 可能接受绝对本地路径并自行传输。如果不得不自行构造
download_url
或
file_id
,说明宿主没有传输桥接机制,应改用下方的 signed-URL 流程。不要虚构文件引用,也不要让文件字节通过模型传输。
Roboflow 服务器会下载文件、组装 ZIP 并上传到存储。客户端无需对文件进行 ZIP 打包,也无需向 GCS 执行 PUT。每次调用可接受图像及标注文件、仅视频项目的 MP4/MOV,或单独一个数据集 ZIP,总计最多 100 个文件 / 1 GB。文件数量较多时,请在宿主和工具限制范围内打包为单个 ZIP。这属于项目媒体导入,而非通用文档存储。
当宿主未在工具的
files
输入中提供文件或无法处理这些文件,但客户端可以对文件进行 ZIP 打包并通过 HTTP PUT 上传到返回的 signed URL 时,使用
image_upload
。仅出现身份验证、校验或服务器错误,并不能说明该桥接机制不可用。如果沙箱无法访问存储,不要反复重试 PUT;请改用宿主辅助传输或项目的 Web 上传器。
两个流程返回
taskId
后(在客户端完成回退流程中的 PUT 之后),轮询
image_upload_status
,直至状态为
completed
或
failed
。即使任务已完成,也要检查
result.uploadErrors
、
annotationErrors
和
warnings
。对于
result.videoIds
中的每一项,还要轮询
video_upload_status
,直至状态为
uploaded
或
failed
;
videosStaged
和 ZIP 完成并不能证明视频已成功导入。即使其他 ZIP 分块失败,也要轮询所有返回的视频 ID。使用每个最终上传成功的
videoId
,而不是去重前的 ID。
result.videoIds
最多包含 100 个 ID,而
result.videosStaged
只统计成功分块中的视频。如果
result.videosStaged
大于返回的 ID 数量,或
result.failedChunks
大于 0,则说明任务并未返回全部视频 ID:轮询已返回的 ID,并将其余视频报告为未验证,同时提供相应计数。
images_search(project_id, media_types=["video"])
搜索的是整个项目,而不仅是本次导入,因此它可能返回旧视频,并遗漏待处理或失败的视频。只有当搜索结果按文件名与上传文件匹配时,才可将其作为辅助证据;绝不能将其视为导入完成的证明。
两个流程采用相同的图像和标注导入行为:
  • 会导入标注。 ZIP 中可以在图像旁放置标注文件。其解析方式与 Web 上传器相同:COCO JSON(
    _annotations.coco.json
    )、YOLO TXT(带
    data.yaml
    或
    classes.txt
    )、Pascal VOC XML、CreateML JSON、CSV 等。标注通过文件名与图像匹配。在分类项目中,图像的父文件夹名即为其类别。
  • 文件夹决定数据划分。 图像或视频路径中任意位置的
    train/
    、
    valid/
    或
    test/
    文件夹(也包括
    training/
    、
    validation/
    、
    val/
    、
    testing/
    )都会设置其 split。其他文件使用
    split
    参数。
  • 带标注的图像会直接进入 Dataset。 无需接受标注任务,除非工作区启用了标注审核(此时图像会等待审核任务)。没有标注的图像会进入一个批次(
    batch_name
    ,默认值为 "Uploaded via API");使用
    annotations_save
    并设置
    add_to_dataset=true
    进行标注。
  • 重复图像仍会获得标注。 已存在于项目中的图像不会再次存储,但 ZIP 中随附的标注会被保存。如果该图像已有标注,则保留原有标注,并在
    annotationErrors
    中将其列为 "Image was already annotated";传入
    annotation_overwrite=true
    可替换它。分类标签始终会被替换。
  • 失败的分块可以安全重试。 ZIP 会按分块处理。如果状态为
    failed
    ,且
    uploadErrors
    中存在类似
    chunk-0
    的条目,则该分块中的图像可能已存储但没有标注。再次上传同一个 ZIP:已存储的图像会作为重复项返回,并补存其缺失的标注。对于首次成功上传的图像,出现 "Image was already annotated" 属于预期情况。

Tags

标签

Tags are free-form labels on images for organization and filtering.
ActionHow
Add during uploadTag selector in upload dialog or via API
Add to existing imagesSelect images -> "Images Selected" -> "Apply tags"
Rename/delete in bulkProject Settings -> Tags -> "Modify Tags"
Filter by tagSearch with
tag:<name>
or use Assign page filter
Use in versions"Filter by Tag" preprocessing step (require/exclude/allow)
标签是图像上的自由格式标签,用于组织和筛选。
操作方式
上传时添加上传对话框中的标签选择器,或通过 API
添加到现有图像选择图像 -> "Images Selected" -> "Apply tags"
批量重命名/删除Project Settings -> Tags -> "Modify Tags"
按标签筛选使用
tag:<name>
搜索,或使用 Assign 页面筛选器
在版本中使用"Filter by Tag" 预处理步骤(require/exclude/allow)

Dataset Search (RoboQL)

数据集搜索(RoboQL)

Search images via the Images page search bar, or with the MCP
images_workspace_search
tool (add
project:<slug>
to stay inside one project). Combine filters with boolean logic.
The MCP
images_search
tool does not parse RoboQL. Its query is semantic text that ranks results without filtering them, so
split:valid
there changes nothing. It filters only through its own parameters (tag, class, Dataset, batch, annotation job). Use
images_workspace_search
for splits and other RoboQL filters.
通过 Images 页面搜索栏搜索图像,或使用 MCP
images_workspace_search
工具(添加
project:<slug>
可将搜索限制在一个项目内)。可以组合多个筛选器并使用布尔逻辑。
images_search
MCP 工具不解析 RoboQL。其查询是语义文本,仅用于对结果进行相关性排序而不执行筛选,因此在该工具中传入
split:valid
不会产生任何作用。它仅通过自身参数(tag、class、Dataset、batch、annotation job)进行筛选。对于 splits 和其他 RoboQL 筛选器,请使用
images_workspace_search
。

Filters

筛选器

FilterExampleDescription
(free text)
person on sidewalk
Semantic search (CLIP-based)
like-image:<ID>
like-image:abc123
Find visually similar images
filename:
filename:*factory*
Filename match (
*
for partial)
tag:
tag:factory
Filter by tag
split:
split:train
Filter by split
job:
job:<JOB_ID>
Filter by annotation job
class:
class:helmet
Has annotation with class
metadata:
metadata:key=value
Filter by user metadata
project:
project:my-project
Filter by project (workspace search)
dataset:
dataset:my-project
Images in the project's Dataset;
project:my-project NOT dataset:my-project
finds the ones not in it yet
sort:
sort:updated
Sort results
min-width:
/
max-width:
min-width:1000
Image dimension filters
min-height:
/
max-height:
max-height:800
Image dimension filters
min-annotations:
/
max-annotations:
max-annotations:1
Annotation count filters
筛选器示例说明
自由文本
person on sidewalk
语义搜索(基于 CLIP)
like-image:<ID>
like-image:abc123
查找视觉上相似的图像
filename:
filename:*factory*
文件名匹配(
*
表示部分匹配)
tag:
tag:factory
按标签筛选
split:
split:train
按 split 筛选
job:
job:<JOB_ID>
按标注任务筛选
class:
class:helmet
包含指定类别标注
metadata:
metadata:key=value
按用户 metadata 筛选
project:
project:my-project
按项目筛选(工作区搜索)
dataset:
dataset:my-project
项目 Dataset 中的图像;
project:my-project NOT dataset:my-project
可查找尚未进入 Dataset 的图像
sort:
sort:updated
排序结果
min-width:
/
max-width:
min-width:1000
图像宽度筛选
min-height:
/
max-height:
max-height:800
图像高度筛选
min-annotations:
/
max-annotations:
max-annotations:1
标注数量筛选

Boolean Logic

布尔逻辑

  • AND
    ,
    OR
    ,
    NOT
    , parentheses:
    class:helmet AND NOT (tag:v1 OR tag:v2)
  • Inverted filter with
    -
    :
    -class:vest
  • Comparison operators on numeric filters:
    >
    ,
    <
    ,
    >=
    ,
    <=
    ,
    =
    (e.g.,
    class:helmet>=3
    )
  • AND
    、
    OR
    、
    NOT
    和括号:
    class:helmet AND NOT (tag:v1 OR tag:v2)
  • 使用
    -
    反向筛选:
    -class:vest
  • 数值筛选器支持比较运算符:
    >
    、
    <
    、
    >=
    、
    <=
    、
    =
    (例如
    class:helmet>=3
    )

Splits (Train / Valid / Test)

Splits(Train / Valid / Test)

Images are assigned to train, valid, or test splits. Splits are rebalanced during version generation (Step 2 in version creation). Augmentations only apply to train split.
图像会被分配到 train、valid 或 test split。创建版本时(版本创建流程的第 2 步)会重新平衡各 split。数据增强仅应用于 train split。

Dataset Versions

数据集版本

A version is a frozen snapshot of the dataset at a point in time. Changes to the project after version creation do not affect existing versions.
版本是某一时间点数据集的冻结快照。版本创建后对项目所做的更改不会影响现有版本。

Version Creation Pipeline

版本创建流程

  1. Source selection — images from the dataset split
  2. Train/Test split — rebalance percentages
  3. Preprocessing — applied to all splits (train + valid + test)
  4. Augmentation — applied only to train split
  5. Generate — creates immutable version
  1. 来源选择 — 来自数据集 split 的图像
  2. 训练/测试划分 — 重新平衡百分比
  3. 预处理 — 应用于所有 split(train + valid + test)
  4. 数据增强 — 仅应用于 train split
  5. 生成 — 创建不可变版本

Preprocessing Options

预处理选项

StepEffect
Auto-OrientStrips EXIF, normalizes orientation
ResizeStretch to / Fit within / Fit (black edges) / Fit (white edges)
GrayscaleConvert RGB to single channel
Auto-Adjust ContrastContrast Stretching / Histogram Equalization / Adaptive (CLAHE)
Isolate ObjectsCrop each bbox into separate image (converts OD to classification)
Static CropCrop all images to fixed region
TileSplit images into NxN grid (default 2x2, helps small object detection)
Dynamic CropCrop images around a specific class
Modify ClassesRemap/omit classes for this version only
Filter NullControl percentage of unannotated images
Filter by TagRequire / Exclude / Allow images by tag
Random SampleSample a percentage of images per split
步骤效果
Auto-Orient移除 EXIF,规范化方向
Resize拉伸至(Stretch to)/ 适配范围内(Fit within)/ 适配(黑边)/ 适配(白边)
Grayscale将 RGB 转换为单通道
Auto-Adjust Contrast对比度拉伸 / 直方图均衡化 / 自适应(CLAHE)
Isolate Objects将每个 bbox 单独裁剪为一张图像(将 OD 转换为分类)
Static Crop将所有图像裁剪到固定区域
Tile将图像拆分为 NxN 网格(默认 2x2,有助于检测小目标)
Dynamic Crop围绕特定类别裁剪图像
Modify Classes仅为此版本重映射或忽略类别
Filter Null控制未标注图像的百分比
Filter by Tag按标签 require / exclude / allow 图像
Random Sample按 split 采样一定百分比的图像

Augmentation Options

数据增强选项

Applied to train images only. Configurable max version size (e.g., 3x = source + 2 augmented copies).
AugmentationImage LevelBBox LevelTier
FlipyesyesBasic
90 deg RotateyesyesBasic
CropyesyesBasic
RotationyesyesBasic
ShearyesyesBasic
GrayscaleyesnoBasic
HueyesnoBasic
SaturationyesnoBasic
BrightnessyesyesBasic
ExposureyesyesBasic
BluryesyesBasic
NoiseyesyesBasic
Camera GainyesyesBasic
Motion BluryesyesBasic
CutoutyesnoEnhanced (paid)
MosaicyesnoEnhanced (paid)
仅应用于训练图像。可配置最大版本大小(例如 3x = 源图像 + 2 份增强副本)。
数据增强图像级BBox 级层级
Flip是是基础
90 deg Rotate是是基础
Crop是是基础
Rotation是是基础
Shear是是基础
Grayscale是否基础
Hue是否基础
Saturation是否基础
Brightness是是基础
Exposure是是基础
Blur是是基础
Noise是是基础
Camera Gain是是基础
Motion Blur是是基础
Cutout是否增强(付费)
Mosaic是否增强(付费)

Dataset Analytics

数据集分析

Available at project sidebar -> "Analytics". Shows:
  • Image count, annotation count, avg image size, median aspect ratio
  • Missing and null annotation counts
  • Class distribution across train/valid/test
  • Image dimension insights (size + aspect ratio distribution)
  • Annotation heatmap (click-drag to filter images by region)
  • Object count histogram (click bars to see matching images)
可在项目侧边栏的 "Analytics" 中查看。显示:
  • 图像数量、标注数量、平均图像尺寸、中位宽高比
  • 缺失和空标注数量
  • train/valid/test 中各类别的分布
  • 图像尺寸洞察(大小 + 宽高比分布)
  • 标注热力图(点击并拖动可按区域筛选图像)
  • 对象数量直方图(点击柱形可查看匹配图像)

Classes

类别

Managed at Project Settings -> Classes.
ActionDescription
RenameType new name in Override column
MergeOverride multiple classes to same name
DeleteCheck Delete checkbox
Lock"Lock Annotation Classes" prevents new class creation
Warning: Class changes at project level affect all images (irreversible). Use version-level "Modify Classes" preprocessing for non-destructive changes.
在 Project Settings -> Classes 中管理。
操作说明
Rename在 Override 列中输入新名称
Merge将多个类别 Override 为同一名称
Delete勾选 Delete 复选框
Lock"Lock Annotation Classes" 可防止创建新类别
警告: 项目级类别更改会影响所有图像(不可逆)。对于非破坏性更改,请使用版本级的 "Modify Classes" 预处理。

Annotation Groups

标注组

Annotation group = the category encompassing all classes in a project. Projects sharing the same annotation group share their class list and annotations.
  • Enable during project creation: "Share image annotations with other projects"
  • Shared annotations: editing in one project affects all linked projects
  • Look for chain-link icon to identify shared images/projects
  • Images shared across projects count only once toward usage
Reusing a group name shares annotations, even with a project in the Trash. The group is just the
annotation
name given at creation. A new project created with a name that another project already uses (trashed projects included) silently joins that group: images uploaded to it that already exist in the workspace come with their annotations, and uploading them with annotations fails with "Image was already annotated". Unless the user wants shared annotations, check the
annotation
values from
projects_list
and
trash_list
before
projects_create
and pick a name no other project uses.
标注组 = 涵盖项目中所有类别的类别集合。共享同一标注组的项目会共享其类别列表和标注。
  • 创建项目时启用:"Share image annotations with other projects"
  • 共享标注:在一个项目中编辑会影响所有关联项目
  • 查找链环图标以识别共享图像或项目
  • 跨项目共享的图像在用量中只计算一次
复用组名称会共享标注,即使该组对应的项目位于回收站中也是如此。该组仅是在创建时给出的
annotation
名称。如果新项目使用了另一个项目已有的名称(包括已删除的项目),它会静默加入该组:上传到该项目且已存在于工作区中的图像会带有标注,而带标注上传这些图像时会失败并提示 "Image was already annotated"。除非用户确实希望共享标注,否则请在调用
projects_create
前检查
projects_list
和
trash_list
中的
annotation
值,并选择一个未被其他项目使用的名称。

Project Folders

项目文件夹

Folders group projects for organization. SSO workspaces can restrict folder access to specific team members.
ActionHow
Create"+ New Folder" from workspace view
Move projectProject menu -> "Move Project"
Delete folderFolder menu -> "Delete" (projects move to workspace root, not deleted)
文件夹用于对项目进行分组。SSO 工作区可以将文件夹访问权限限制给特定团队成员。
操作方式
Create从工作区视图点击 "+ New Folder"
Move project项目菜单 -> "Move Project"
Delete folder文件夹菜单 -> "Delete"(项目会移至工作区根目录,不会被删除)

Export Formats

导出格式

Versions can be exported as
.zip
download or
curl
command. 40+ formats supported including COCO, YOLO, Pascal VOC, TFRecord, and more. Full list at
roboflow.com/formats
.
Export via Python SDK:
python
project.version(1).download("yolov8")
版本可以导出为
.zip
下载或
curl
命令。支持 40 多种格式,包括 COCO、YOLO、Pascal VOC、TFRecord 等。完整列表见
roboflow.com/formats
。
通过 Python SDK 导出:
python
project.version(1).download("yolov8")

Project type and license

项目类型和许可证

Don't guess a project's type or license. Confirm both with the user before calling
projects_create
.
不要猜测项目类型或许可证。调用
projects_create
前,请与用户确认两者。

MCP Tools Available

可用的 MCP 工具

ToolPurpose
projects_create
Create a new project (specify type and an unused annotation group unless sharing is intended)
projects_list
/
projects_get
List or get project details
images_search
Semantic search inside one project, with tag, class, Dataset, batch and job filters (no RoboQL)
images_workspace_search
RoboQL search across the workspace, e.g.
project:my-project split:valid
image_upload_files
/
image_upload
/
image_upload_status
Prefer host-assisted image uploads; use client ZIP + signed PUT when needed; poll the import task
versions_generate
Generate a dataset version with preprocessing/augmentation
versions_get
Inspect a version
versions_export
Export a version in a given format
工具用途
projects_create
创建新项目(指定类型和未使用的标注组,除非有意共享)
projects_list
/
projects_get
列出项目或获取项目详情
images_search
在单个项目内进行语义搜索,可按 tag、class、Dataset、batch 和 job 筛选(不支持 RoboQL)
images_workspace_search
在工作区中进行 RoboQL 搜索,例如
project:my-project split:valid
image_upload_files
/
image_upload
/
image_upload_status
优先使用宿主辅助图像上传;必要时使用客户端 ZIP + signed PUT;轮询导入任务
versions_generate
使用预处理和数据增强生成数据集版本
versions_get
检查版本
versions_export
以指定格式导出版本

Related Pages

相关页面

  • skill://roboflow/roboflow-data-management/labeling.md
    — annotation tools, AI labeling, Label Assist, Smart Polygon, Auto Label, annotation jobs
  • skill://roboflow/roboflow-cloud-storage/SKILL.md
    — mirror an S3/GCS bucket into the workspace (credentials, datasources, glob rules, scheduled sync)
  • skill://roboflow/roboflow-data-management/labeling.md
    — 标注工具、AI labeling、Label Assist、Smart Polygon、Auto Label、标注任务
  • skill://roboflow/roboflow-cloud-storage/SKILL.md
    — 将 S3/GCS bucket 镜像到工作区(凭证、数据源、glob 规则、定时同步)