High-quality data annotation, done right.
LabelRight turns raw text, images and audio into structured, accurately labelled datasets for training, evaluating and improving AI models.
Launch offer Try LabelRight: get your first dataset sample annotated for free.
Text · entities and sentiment
Harbour Freight CoORG confirmed the delivery to CardiffPLACE will arrive on 14 MarchDATE, two days late.
Sentiment negative · Intent delivery update
Image · bounding boxes
Audio · speakers and transcript
Speaker A “Can you confirm the order number?” Speaker B “Yes, it ends four two seven.”
Your first dataset sample, annotated for free.
We are introducing LabelRight to the market and want businesses to experience our approach to data annotation before committing to a larger project.
- Up to 100 straightforward annotations or 2 hours of work, whichever is reached first.
- One free trial project for each company.
- For selected small datasets. We confirm the scope before we start.
- A basic quality review is included.
- You receive the annotated sample in an export format we agree with you.
- No obligation to buy anything afterwards.
- Larger or more complex jobs receive a custom quotation.
The trial scope is subject to our approval. For time-intensive tasks, such as audio or detailed image outlines, we may offer a smaller sample.
Who labels your data
- UK-based. Every annotator lives and works in the UK. Your data is labelled here, not sent abroad.
- DBS-checked. Every annotator has passed a DBS (criminal record) check before they see customer data.
- Fluent English. Every annotator works in fluent English, which matters for text, guidelines and edge cases.
What we label
Each service shows its status. “Available” means we have run it through our full workflow, from upload to export. “By arrangement” means we scope it with you first.
Text annotation
Labels for language data: from simple categories to judgements on model output.
- Text classification and categorisationAvailable
- Named entity recognitionAvailable
- Sentiment analysisAvailable
- Document labellingAvailable
- Intent classificationAvailable
- LLM response evaluationAvailable
- Human preference and response rankingAvailable
Image annotation
Labels for computer vision datasets.
- Image classificationAvailable
- Object detection and bounding boxesAvailable
- Image segmentationBy arrangement
- Image taggingAvailable
Audio annotation
Labels and text for speech and sound recordings.
- Speech transcriptionBy arrangement
- Speaker identification and diarisationBy arrangement
- Audio classificationBy arrangement
Dataset preparation
The work around the labels that decides if a dataset is usable.
- Cleaning and organising datasetsAvailable
- Creating annotation guidelinesAvailable
- Data formatting and exportAvailable
- Quality control and annotation reviewAvailable
- AI-assisted pre-labelling with human verificationBy arrangement
From enquiry to labelled dataset
- 1Enquiry
- 2Assessment
- 3Secure dataset submission
- 4Project set-up
- 5Annotation and quality control
- 6Delivery
- 7Completion
How we handle your data
- You send data through a private upload link made for your project. There is no public upload.
- Each file is scanned for malware and stored encrypted.
- Your dataset is kept apart from every other customer's data.
- We do not send your data to outside AI services unless you ask us to in writing.
- We delete your data on a date we agree with you, and we record the deletion.
Tell us about your dataset.
Start with a free sample, or ask for a quotation for a larger project.