I've been waiting to play around with Google's new Gemini 3.0 and decided to show, not tell. I created a full idle "cookie clicker" game in <10m in AI Studio. You can upgrade your equipment, your team, or raise new money from Y Bombinator and more to get the full Silicon Valley experience.
Ivan chiming in from Datasaur here. As jordn pointed out, Datasaur does view itself as a full labeling platform, which encompasses an optimized labeling interface, a workforce management tool in addition to intelligence and active learning. Unlike Humanloop, we are focused solely on the labeling step of the process and do not offer a trained model at the end of the process. Our users have separate pipelines for this. Thanks for the question!
Hi IanCal -
Really appreciate you taking the time to test Datasaur out and provide this feedback! Responses below:
- We took the risk of launching before our tutorial is in place to get feedback, so I do apologize for lack of clarity for first time users.
- Good point, we'll try improve error messages for users.
- We've been taking import formats on a case-by-case basis, but we're looking to improve/expand our import/export format flexibility further.
- This is probably one of the highest priority issues for us to fix - making sure expected format and input is clear to the user.
- Happy to send you additional sample files and instructions on how to get autolabelling working. We can accept your own models/api endpoints as well!
- Hotkeys for labeling should work - this bug seems odd. What kind of project is this for?
- Search and label should work as soon as labels have been uploaded, sorry that it wasn't working for you.
Happy to discuss any of this in further detail. Thanks again for leaving such comprehensive thoughts!
We're fans of brat and are hoping to take it to the next level. We devoted many hours to ensuring a highly performant web app on the latest modern tech stack. Additionally we're expanding beyond a labeling interface to include semi-automated annotation, as well as team management capabilities.
Both can be described as semi-automated annotation, but we use different approaches! Datasaur allows you to plug in any pre-existing model to pre-label or validate your labels. One such model we integrate with is spaCy, so we're certainly fans of the Prodigy/Explosion team.
Hi there - you may choose to upload labels as a text file or create your own. I'd be curious to hear more about your use case in batch-applying labels. I'll follow up offline (well, via email).
Always happy to hear people complimenting the logo. I've been told it's not professional enough, but I really wanted our site to have some personality. Thanks!
So here's the secret. Our awesome designer actually put a lot of hard work into putting this together. 1 week later a friend told me about https://www.landen.co/ and I wish I had used this to save us some time (don't know the team, just a fan of the product).
Done and shipped! :D
One of our extensions allows you to plug in an API - either use your own model, or an integration with an open-source project like spaCy to apply labels.
Thanks - good to see so many people concerned about privacy here. I consider a privacy a top-level priority at Datasaur - all data is fully encrypted. Our employees will never be able to see or access any customer data. We already work with a bank and cleared their security bar :)
Certainly. Our philosophy is to complement human wisdom with computer precision. Humans may often be labeling for 8 hours a day and may get fatigued. So if Starbucks has been labeled as a cafe 35x in a document and as a person 2x, we can flag this and ask "hey, are you sure you wanted to label this as a person?". Or if we know for a fact Canada is a country, but it's labeled as an animal in a document, we can raise a flag as well. This won't work for everything, but we think it can help with quality assurance.
That's right! Scale probably has some awesome internal tools that help them label faster. Datasaur wants to make those same optimizations available to anyone with their own labelers.
Apologies for the delay! There is some overlap with what Prodigy works on and I'm a big fan of what they're working on. We cover some additional use cases (like coreference parsing) and additionally help with managing teams of labelers. We're complementary in many regards. Happy to discuss further, based on your labeling needs.
We offer both a hosted service on AWS as well as an on-prem solution if needed. We can even choose to host on the cloud provider of your choice - happy to work with you on this!
Great question!
Data privacy is a top-level priority for us. We actually offer both a cloud-based and on-prem solution. One of our clients needed a fully on-prem, air-gapped (no connection to internet) option. Many are choosing to use us because they can't send their data to outsourced, external parties.