Hasty uses credit-based pricing. This can be confusing to some users, leading to the question...
Credits are our own in-tool currency. The concept might be familiar to annotators who also dabble in gaming. For others, it can best be explained as a way to pay for what you use in the tool.
Essentially, whatever you want. What's important to know with Hasty is that we never charge you for doing manual tasks like annotating completely by hand. We also don't charge you per user or seat. We are engineers ourselves and we like to pay only when we get some utility back. This is at the core of our pricing.
You get charged when you use any automated or semi-automated annotation features, any automated QA features, and when you train models using Hasty. Essentially, we charge you when we:
Deliver value to you that either saves you time...
Give you better results than you would get by doing it by hand...
Or when you do an action that has a high cost to the company (like training a machine learning model)
Additionally, we charge a small hosting fee for any data uploaded to Hasty. We do so to prevent some spammy behavior that we saw at the start of our existence as a company.
Yes, we do.
Cost in credits per unit
Tagger or Attribute Assistant
Object Detection Assistant
Box to Instance Assistant
Instance Segmentation Assistant
Semantic Segmentation Assistant
As per assistant used
Error Finder (training of model)
Per Checked Annotation
Error Finder Processing
Per Checked Annotation
Absolutely. With your 3000 credits, you can:
When doing image classification:
Create 3000 image tags automatically OR
Create ca. 1900 image tags automatically and do one AI quality control run on those OR
Create ca. 1600 image tags automatically, do one AI quality control run, and train one AI model
When doing object detection:
Create 1500 bounding boxes automatically OR
Create 1150 bounding boxes automatically and do one AI quality control run on those OR
Create 950 bounding boxes automatically, do one AI quality control run, and train one AI model
...once again, every month
When doing instance/semantic segmentation:
Create 1000 segmentation masks/polygons automatically OR
Create 850 segmentation masks/polygons automatically and do one AI quality control run on those OR
Create 700 segmentation masks/polygons automatically, do one AI quality control run, and train one AI model
...and yes, this you can do every month too
You have two options. Either, you wait to next month. Or, you upgrade to a paid plan.
Yes - just email Herbert at [email protected] and tell us about what you are researching or studying and we will give you a discount.
Same question - but now I am from an NGO - can I still get a discount?
Yes. With NGOs, we give discounts on a case-by-case basis factoring in independent reports on governance and the like.
I want to upgrade to a paid plan but have problems with my payment method - what do I do?
Contact Herbert (email above) and he will help you.
Not in-app, but once again, Herbert can help you with this.
What happens if I buy a paid plan but don't use up all my credits?
Credits have an expiry date of 6 months - so if you don't use up your credits from the first month during the first month, you have five additional months to catch up.
Why do you use credits and not Dollars/Euros/Yen etc.?
Purely for accounting purposes. If we metered you in an actual currency, we would have to factor in currency fluctuations into our accounting - in essence, this would mean we would have to spend all our money on accountants. Accountants are generally friendly people but we prefer to not give them all our money.
What is a credit worth?
On a free plan - nothing. On a paid plan it depends on the plan. The price per credit gets lower with larger plans.