Skip to content
SPECIFICTION
Get listed

MakeSense.ai

makesense.ai

Free plan Open source WebWindowsmacOSLinuxSelf-hosted
Visit makesense.ai
Spec score 8.6 #3 of 28 in Image Annotation Software

What it is

MakeSense.ai is a browser-based tool for labeling photos and preparing datasets for computer vision projects. It supports point, line, rectangle, polygon, and label annotations, with model-assisted labeling through YOLOv5, SSD pretrained on COCO, and PoseNet. Users can load their own YOLOv5 models exported to TensorFlow.js format. Export options include CSV, YOLO, VOC XML, VGG JSON, and COCO JSON; imports include YOLO, VOC XML, and COCO JSON. The available format combinations vary by annotation type, and some are not yet supported. The project says images are not sent to a server and AI features run using TensorFlow.js. It works in a browser across operating systems, and documented local and Docker setups are available; Docker serves the application on localhost:3000. The project describes it as suitable for small computer vision deep-learning projects and dataset preparation. It was founded in 2019, is licensed under GPL-3.0, and credits Piotr Skalski as author and copyright holder. The Free plan costs nothing.

Spec sheet

What it does
MakeSense.ai is a browser tool for labeling photos and preparing computer-vision datasets.Read 21 Oct 2026 · github.com
Intended users
The project describes the tool as suitable for small computer vision deep learning projects and dataset preparation.Read 7 Oct 2026 · github.com
Privacy
The project says images are not stored because they are not sent to a server.Read 21 Oct 2026 · github.com
Founded
2019Read 28 Sep 2026 · makesense.ai
Maker
The repository credits Piotr Skalski as author and copyright holder, with a 2019 project copyright.Read 21 Oct 2026 · github.com
AI models
Its listed AI integrations include YOLOv5, SSD pretrained on COCO, and PoseNet.Read 7 Oct 2026 · github.com
Annotation features
The tool supports point, line, rectangle, polygon, and label annotations.Read 7 Oct 2026 · github.com
API and SDK documentation
The public documentation presents no product API or SDK; its documented extensibility is local code, model loading and inference-server connection.Read 21 Oct 2026 · skalskip.github.io
Cross-platform
The project describes itself as cross-platform because it runs in a browser across operating systems.Read 21 Oct 2026 · github.com
Cross-platform use
The project says it works in a browser across operating systems and requires no complicated installation.Read 7 Oct 2026 · github.com
Custom models
Users can load their own YOLOv5 models exported to TensorFlow.js format.Read 7 Oct 2026 · github.com
Docker self-hosting
The repository includes Docker instructions for running the application locally.Read 21 Oct 2026 · github.com
Documentation
The public documentation contains separate pages for imports, exports, shortcuts and local setup.Read 21 Oct 2026 · skalskip.github.io
Export formats
Supported exports include CSV, YOLO, VOC XML, VGG JSON and COCO JSON, depending on annotation type.Read 21 Oct 2026 · skalskip.github.io
Image classification
Image-level labels are available for classification tags.Read 21 Oct 2026 · deepwiki.com
Implementation stack
The application is written in TypeScript and uses React with Redux.Read 21 Oct 2026 · github.com
Import formats
Supported imports include YOLO and COCO JSON for rectangles and COCO JSON for polygons; other combinations are marked unsupported or not applicable.Read 21 Oct 2026 · skalskip.github.io
Independent praise: privacy
CheckThat.ai characterizes the tool as free, open-source, browser-based and processing data locally for privacy.Read 21 Oct 2026 · checkthat.ai
Independent use
Researchers have used MakeSense.ai in scientific image-labeling work, including X-ray scattering datasets.Read 21 Oct 2026 · journals.iucr.org
Latest release
The releases page lists version 1.11.0-alpha dated 20 December 2022 as the latest release shown.Read 21 Oct 2026 · github.com
Licence
The code is released under the GNU GPL-3.0 licence.Read 21 Oct 2026 · github.com
Limit: batch workflows
The same paper reports that MakeSense.ai lacks batch labeling and similarity-based annotation suggestions.Read 21 Oct 2026 · journals.iucr.org
Limit: image adjustment
The paper reports no built-in brightness or contrast adjustment tools.Read 21 Oct 2026 · journals.iucr.org
Limit: model accuracy
A 2021 CEUR paper describes the neural-network choices as limited and reports low object-detection accuracy in its comparison.Read 21 Oct 2026 · ceur-ws.org
Limit: thumbnail information
The paper reports that thumbnails show whether an image is labeled but not its specific labels or label totals.Read 21 Oct 2026 · journals.iucr.org
Line annotation
Lines support edge or boundary annotations.Read 21 Oct 2026 · deepwiki.com
Local deployment
The repository documents cloning the code, installing dependencies with npm and serving it at localhost:3000.Read 21 Oct 2026 · skalskip.github.io
Local hosting
The project documents local setup and a Docker setup that serves the application on localhost:3000.Read 7 Oct 2026 · github.com
No installation
The web application can be used by opening the site in a browser without a complicated installation.Read 21 Oct 2026 · github.com
Notable limits
Its format matrix marks some annotation and import/export combinations as not yet supported.Read 7 Oct 2026 · github.com
Point annotation
Points are available for keypoint or landmark labeling.Read 21 Oct 2026 · deepwiki.com
Polygon annotation
Polygons support segmentation annotations.Read 21 Oct 2026 · deepwiki.com
PoseNet integration
PoseNet is included for estimating human poses in images or video.Read 21 Oct 2026 · github.com
Praise: usability
A 2025 Journal of Applied Crystallography paper describes the interface as intuitive and says its one-by-one manual labeling speed was comparable to the authors’ pipeline.Read 21 Oct 2026 · journals.iucr.org
Project license
The project is licensed under GPL-3.0.Read 7 Oct 2026 · github.com
Rectangle annotation
Rectangles provide bounding-box annotations for object detection.Read 21 Oct 2026 · deepwiki.com
Review coverage
CheckThat.ai reports minimal presence for MakeSense.ai on major business-software review platforms and no enterprise documentation.Read 21 Oct 2026 · checkthat.ai
Roboflow integration
Release notes describe connecting to a Roboflow inference server for annotation assistance.Read 21 Oct 2026 · github.com
SSD integration
A COCO-pretrained SSD model can suggest bounding boxes and labels locally.Read 21 Oct 2026 · github.com
System requirements
The local-setup documentation specifies npm 6.x.x and Node.js v11.x.x.Read 21 Oct 2026 · skalskip.github.io
System-requirement inconsistency
The repository README instead specifies npm 8.x.x and Node.js v16.x.x, so the published setup requirements differ.Read 21 Oct 2026 · github.com
TensorFlow.js
The AI functions use TensorFlow.js in the browser.Read 21 Oct 2026 · github.com
YOLOv5 integration
Users can run pretrained or custom YOLOv5 models exported to TensorFlow.js format.Read 21 Oct 2026 · github.com

Plans and prices

Free Free browser photo labeling · no image upload to server Read 21 Oct 2026 · github.com

Every MakeSense.ai plan and price →

Strong points and limits

Strong points

  • Supports five annotation types.
  • Includes YOLOv5, SSD, and PoseNet integrations.
  • Accepts custom YOLOv5 models in TensorFlow.js format.
  • Exports to five listed formats.
  • Free plan costs nothing.

Limits

  • Some annotation and format combinations are unsupported.
  • Custom models must be YOLOv5 models exported to TensorFlow.js.

Who it suits

It suits people preparing datasets for small computer vision projects who need several annotation types and export formats. Browser access or local hosting may suit different setup preferences.

MakeSense.ai in depth

Read the full write-up

MakeSense.ai is a free image-labeling tool for people preparing computer-vision datasets, especially for small projects that can run in a browser or be hosted locally. Its appeal is straightforward: several annotation types and model-assisted labeling without a paid plan; its key constraint is format compatibility.

Overview

MakeSense.ai concentrates on annotating photos and preparing datasets rather than covering a broader computer-vision workflow. It suits individual users and small project teams who want a focused labeling tool, but projects tied to a specific annotation and file-format combination should check the format matrix before committing.

The project says images are not sent to a server or stored, and that its AI features run with TensorFlow.js. Browser use avoids a complicated installation, while documented local and Docker setups give users a self-hosting option; Docker serves the app at localhost:3000. That flexibility is useful, though the tool's emphasis remains on small dataset-preparation projects.

Key features

Annotation and model assistance

Annotations include points, lines, rectangles, polygons, and labels. Model-assisted labeling supports YOLOv5, SSD pretrained on COCO, and PoseNet; users may also load their own YOLOv5 models if exported to TensorFlow.js format. This gives small projects several ways to speed up labeling, but custom-model support is specifically tied to that export format.

Dataset formats

Exports include CSV, YOLO, VOC XML, VGG JSON, and COCO JSON; imports include YOLO, VOC XML, and COCO JSON. The format matrix marks some annotation and import/export combinations as unsupported. MakeSense.ai is a practical fit when the project's required path is covered, not when a pipeline depends on an unsupported combination.

Pricing

The Free plan costs nothing and includes browser-based photo labeling, model-assisted labeling, and dataset export. It is a strong match for individuals and small projects that need the supported annotation and format options, with no paid plan required for those uses. No paid tier is described, so teams whose needs exceed the tool's supported formats should consider another product rather than expect a larger MakeSense.ai plan.

Platforms

MakeSense.ai runs on the web and supports Linux, macOS, and Windows through a browser. The project also documents self-hosting, including a Docker setup served at localhost:3000. Browser-based use avoids a complicated installation; self-hosting is available for users who prefer to run the application locally.

Who it's for

Choose MakeSense.ai for small computer-vision deep-learning projects and dataset preparation when its annotation types and format combinations match your workflow. Its free plan, browser access, and local hosting options make it especially suitable for individuals and small teams. Look elsewhere if an unsupported import/export combination is essential or if your work calls for a broader workflow than image labeling.

Pros and cons

Pros

Free to use: the free plan provides browser-based photo labeling without a paid tier.

Several annotation types: point, line, rectangle, polygon, and label annotations serve different dataset-preparation needs.

Model assistance and custom YOLOv5 support: built-in options include YOLOv5, SSD pretrained on COCO, and PoseNet, with custom YOLOv5 models supported in TensorFlow.js format.

Browser use and self-hosting: it works across operating systems in a browser and offers local and Docker setup options.

Privacy-oriented approach: the project says images are not sent to a server or stored.

Cons

Some format combinations are unsupported: a required annotation and import/export path may not fit the format matrix.

Custom model format is constrained: user-supplied YOLOv5 models must be exported to TensorFlow.js format.

Focused scope: the project is aimed at small dataset-preparation work, so it may not suit users seeking a wider computer-vision workflow.

Alternatives

Image Annotation Software, AI Image Annotation Tools, and Data Labeling Software are broader directories for comparing tools in these categories.

Label Studio is worth considering if you want a freemium option with a free trial and API support alongside web and self-hosted use.

Roboflow may suit projects that need Android or iOS access; its Free Tier is free and includes 10 credits a month.

Segments.ai is an alternative for academic users: its free Academic license is for non-commercial research and requires a license request.

Supervisely may fit a small team looking for a freemium platform with a Community plan that includes 5 GB of storage, 10,000 files, and 2 members.

VGG Image Annotator is another free option, released under the two-clause BSD license, with no paid tiers or usage limits stated.

CVAT may suit users who want an open-source Community plan for personal use and small teams, or its online free plan for a limited project.

Labelme is an alternative for users who prefer a desktop app and SAM2/SAM3 annotation; its Starter plan costs $49 as a one-time purchase and includes one year of updates.

Pixano is another free option.

Verdict

MakeSense.ai is a sensible choice for individuals and small teams preparing image datasets who want browser-based or self-hosted labeling, model assistance, and no-cost use. Its strongest reason to choose it is that focused toolkit at free; its strongest reason to look elsewhere is the format matrix, which can make it unsuitable for projects with fixed, unsupported import/export requirements.

Verdict

MakeSense.ai combines browser-based image annotation with model-assisted labeling and multiple dataset formats. Check its format matrix against the annotation and import/export combinations your project needs.

Where Specifiction places it

Alternatives to MakeSense.aiAll alternatives →