Picsellia
Toulouse computer vision MLOps platform with annotation, review and SAM-assisted labeling built in
Quick Overview
| Company | Picsellia SAS |
|---|---|
| Category | Data Labeling Platforms |
| Headquarters | Toulouse, France |
| Founded | 2020 |
| EU Presence | Yes - France (EU) |
| Data Location | EU and US data centres (SaaS); your own infrastructure for hybrid and on-premise |
| Open Source | No |
| GDPR Compliant | Yes |
| Self-Hosting | Yes, on-premise with Kubernetes or Docker Compose, air-gapped supported |
| Main Features | Bounding box, polygon, segmentation, keypoint and polyline labeling, SAM and Grounding DINO pre-labeling, review workflow, dataset versioning, training, deployment and monitoring |
| Pricing | Data Engine from 1,000 euro per month; labeler seats 10 euro per month each; annotation services 9.60 euro per hour |
| Best For | Industrial computer vision teams that want labeling and MLOps in one platform |
| Replaces | Scale AI, Labelbox |
Detailed Review
Picsellia treats annotation as one stage in a computer vision pipeline that also covers data management, training, deployment and monitoring. Its labeling tool supports bounding boxes, polygons, segmentation masks, keypoints and polylines on images, video, satellite and medical scans.
The privacy policy names Picsellia SAS in Toulouse, France, as data controller and the French authority CNIL as the complaint route.
AI-assisted labeling with review
The tool offers SAM, DINOv2 and Grounding DINO for pre-labeling and automated processings for pre-annotation and auto-tagging. Every annotation goes through a review flow in which it can be accepted, rejected or sent back, and reviewed annotations are locked.
Deployment options
The enterprise page lists managed SaaS with EU and US data centres, a hybrid mode in which you keep storage and compute and Picsellia runs the control plane, and an on-premise install on Kubernetes or Docker Compose with air-gapped environments supported.
Prices on the page
The pricing page lists the Data Engine module at 1,000 euro per month or 12,000 euro per year with three seats, extra users at 500 euro per month, labeler seats at 10 euro per month and annotation services at 9.60 euro per hour.
Limitations to Consider
Picsellia is a computer vision platform, so it does not label text or documents. The SaaS option uses EU and US data centres, which means data residency has to be requested or handled through hybrid or on-premise deployment. The entry price of 1,000 euro per month for the Data Engine module is high for a single annotator, for whom Supervisely or SentiSight.ai are cheaper.
Who Should Use Picsellia
Industrial and enterprise computer vision teams that want annotation, dataset versioning and model monitoring in one European platform.
Pros and Cons
Pros
- French contracting entity (Picsellia SAS)
- Labeling, training and monitoring in one product
- SAM-assisted pre-labeling and review workflow
- On-premise and air-gapped deployment
- Euro prices published
Cons
- Vision only, no text labeling
- SaaS uses EU and US data centres
- High entry price per module
- Extra users cost 500 euro per month on Data Engine
Alternatives to Picsellia
Other European data labeling platforms worth comparing before you decide:
Kili Technology
France · Expert-in-the-loop labeling and AI evaluation platform
Kognic
Sweden · Annotation platform and workforce for autonomous driving data
Prodigy
Germany · Scriptable self-hosted annotation tool from the spaCy makers
Supervisely
Estonia · Labeling, data management and model training in one platform
See all European data labeling platforms or compare them directly against Scale AI on our alternatives page.
Frequently Asked Questions
The enterprise page lists SaaS with EU and US data centres, and hybrid or on-premise options for data that must stay in your infrastructure.
The Data Engine module is 1,000 euro per month or 12,000 euro per year with three seats, plus 10 euro per month per labeler seat.
The labeling tool page lists images, videos, satellite imagery, medical scans and multispectral data, so it is a vision tool.