The Text Annotation Tool to Train AI
Turn text into intelligence. Easy.START NOW Discover our PDF annotation tool!
WHAT ARE YOU LOOKING FOR?
Get relevant insights from text, automatically
Discover patterns, identify challenges, realize solutions
- > Analyze user feedback and design specific actions for improvement
- > Enrich your unstructured text with metadata and make it searchable
- > Find out bid opportunities in your industry
Train your own AI
Generate training data for your own ML methods
tagtog is as its core a NLP text annotation tool. Create labeled datasets:
- > Manage your team to annotate text manually (import pre-annotated data too)
- > Leverage machine-learning models to work at scale and semi-supervised
- > Find out biases in your data, and the quality of your annotations
HOW DOES IT WORK?
Upload your own files (PDFs, plain text, CSVs, source code files, ...) or point to external resources like web URLs.
Annotate manually or automatically
Intuitively enrich text: annotate entities and disambiguate (e.g. company products, intents, etc.), classify documents and entities, draw relationships (e.g. diseases caused by mutations).
tagtog learns from your annotations to annotate relevant information automatically. If you prefer, you can plug your own ML model in and use the feedback to train it.
We have trained models ready to extract named entities automatically (e.g. vehicle parts, genetic mutations, etc.).
Export the annotations in various formats using the API or the web interface. Use our search engine to discover actionable insights and make smarter decisions. That’s it!
No Coding or Data Engineering skills required
Democratize Text Analytics. You don’t need to code or juggle with data to use tagtog. Use the intuitive web interface to create high-quality training data by just annotating.
Team Collaboration and Quality management
Invite other users to annotate text and create an annotated corpus. Define guidelines and roles at any moment. Track annotation progress and quality. You can distribute tasks automatically among users based on your quality requirements. More information
Track quality and compare the performance of the different annotators using the inter-annotator agreement (IAA).
The best Text Annotation Tool
Classify documents and entities manually or automatically. Annotate and disambiguate entities (e.g. company products, intents, etc. ), draw relations (e.g. diseases caused by mutations). More information
Do you want to train your own algorithms? Import your predictions, correct them in the annotation tool, and feed them back.
Machine Learning with people in the loop
Plug your ML model and let your team of subject-matter experts provides feedback on the predictions for a continuous training. Improve quickly the quality of your training data and the accuracy of your machines. More information
On our secure Cloud or OnPremises
On the Cloud, there is nothing to install, no servers to worry about: start right now. On-premises, run tagtog as a docker image in your own infrastructure, SSO integration, Internet access is not required. In both cases, just use your favorite browser. More information
Work directly with your documents, not only plain text. Annotate natively over PDF or import the text from TXT files, HTML, CSV, PDF files, source code files, Markdown, etc.
English, Spanish, Hindi, Bengali, French, Chinese, Japanese, Arabic, Swedish, Dutch, etc. Any language. Unicode support. Left to Right and Right to Left.
ML and dictionary annotations
tagtog uses ML to learn from your annotations and generate similar annotations automatically. In addition, you can upload already-annotated documents or term dictionaries. Build high-quality training data in hours.
Integrate tagtog within your existing workflow. Use the API to upload text, retrieve the results or manage folders. You can also use it to search across your text collection. More information
Organize your text and documents in different folders and levels for a better organization. For example, separate test and production data.
Make the most of your data, quick. Annotate overlapping entities or contained within others. Don't miss important information. More information
Search in text collections not by keyword, but by concept (e.g. find all vehicle technical reports that are related to engine failures). More information
Manage disambiguation. tagtog determines the identity of the annotations assigning unique ids from standard databases such as UniProt or Wikipedia.
You can also upload your own dictionaries to map the annotations to your unique internal references (e.g. product ref). More information
Work with external text sources (e.g. PubMed) or your own files. Process millions of text items with ease.
WHY THEY LOVE IT
We were looking for a way, not only of annotating aspects of the historical documents we work with in order to later extract information from these, but also to do so in an expedite manner and with people that is no expert in NLP or ML. We found tagtog online and it was love at first sight. It was the platform we were looking for.
tagtog has been instrumental in our labeling efforts. We have a complex dataset and several sets of class labels. tagtog allowed our non-technical users to annotate large documents with ease, and allowed our data team to process their work using a sophisticated API. We are extremely satisfied with our investment.
tagtog allows us to have a fine-tuned control over tagging in a collaborative environment. AND the tagtog team is very responsive to our questions and new feature requests.
NEW PDF ANNOTATION TOOL
Annotate over the PDF
Annotate directly over the native PDF layout, annotators love it! 💖
Train your models easily
Import/Export annotations using text offsets or coordinates, tagtog gives you also the text contained in the PDF to facilitate the processing and generation of annotations.
Complete PDF viewer
Navigate the document just by scrolling, zoom, pan (hand tool) or search across the document.
Annotate any text in the PDF
Annotate text in figures, tables, pictures, etc.
It allowed us very easily to break down the different entities into selectable categories from the document, directly on PDF. Getting back both the text as well as its relative text positions, this was greatly helpful for us to create an NLP model with Machine Learning.
Creating training data is expensive; you want the right tool.
Easily integrate it with your current workflow, software, and team.
Engage subject-matter experts. Manage NLP projects efficiently.
Create an account and start making sense of your data now
We are a startup based in the charming cities of Munich and Gdańsk.
We love hearing your requests, feedback, and questions:
(we speak English 🇬🇧, Español 🇪🇸, and Polski 🇵🇱!)
(Better yet, let's meet in person!)