Scale AI Documentation

Welcome to Scale! You'll find comprehensive guides and API references to help you start working with Scale AI as quickly as possible. Let's jump in!

Documentation    API Reference

Introduction to Scale

Scale AI’s mission is to accelerate the development of AI applications. To enable teams to make faster progress, we began with data - the foundation of all AI applications. Scale AI turns raw data into high-quality training data by combining machine learning powered pre-labeling and active tooling with varying levels and types of human review.

We provide two tiers for our customers to use our platform:

On Demand
Enterprise

Our On-Demand model requires no long-term commitment and is a pay-as-you-go model. This is ideal for projects with lower volumes with best-effort quality and throughput. To get started, simply create an account.

Our Enterprise model has annual commitments with volume-based rates. This is a deal for strategic AI initiatives with guaranteed turnaround times and quality. To speak with a sales representative, please reach out to us.

More information on both models can be found on our pricing page.

Ready to get started? Follow the 4 steps below and you’ll be well on your way to building a robust training dataset for your next AI application.

1. Choose your Labeling Task

Scale offers high quality training and validation data for AI applications. Both On-Demand and Enterprise customers have access to:

Furthermore, our Enterprise customers have access to:

2. Design your Task

Writing clear and concise instructions is crucial to receiving high-quality training data with Scale.

We strongly encourage you to use the following template to get started.

Clear and concise instructions should include:

  • Step-by-step directions
  • Rules and example images (i.e. minimum label size, when to not label an object)
  • Label classes and attributes / fields with definitions and example images
  • Screenshots of correct and incorrect tasks for comparison

To embed your google docs instructions into your tasks:

  1. In your Google Docs, click File →Publish To Web → Embed (in the window that appears)
  2. Copy the iFrame tag that’s provided: <iframe src="[YOUR_CUSTOM_DOC_LINK]"></iframe>
  3. Paste the iFrame tag in the instruction parameter on the task, or paste it into your project instructions in the UI.

We have created detailed example instructions and guides on our instructions page.

3. Set up your Data Pipeline

Once you have explored your specified task type and designed your task, write your REST API calls following our API documentation. We have examples in cURL, JavaScript, Python, and Ruby. You can submit your API calls via Terminal or a 3rd-party software like Postman.

📘

Test/Live Tasks

Our customer dashboard allows you to toggle between:

  • Test Tasks : Mock tasks that are meant for testing purposes only
  • Live Tasks : Paid tasks that are meant for labeling

Upon account creation, you'll have two API keys, one for testing and one for live calls that will actually be annotated by a human.

4. Iterate and Learn

🚧

Use small batches (5-10 tasks) for new projects

We strongly recommend you send in your data in batches. Your first batch should be small (5-10 tasks). Once you feel comfortable with the quality, please proceed with larger batches

After you submit your first batch, you can explore the web application:

  • Overview tab: View a summary of your submissions over time and the status for each task type.
  • Tasks tab: View a more granular list of your submitted tasks, filterable by status, project, and batch.
  • Quality tab: Audit a random sample of tasks for quality and view stats for tasks you’ve reviewed

Submitting your projects to Scale should be an iterative process. After your first batch, please summarize common errors or specify new requirements in the instructions document. Once you have satisfied with the quality, feel free to increase the batch volume

Need Help?

If you have any questions, feel free to email us at [email protected].

If you are interested in learning more about our Enterprise engagements, get in touch with us.

Updated about 23 hours ago



Introduction to Scale


Suggested Edits are limited on API Reference Pages

You can only suggest edits to Markdown body content, but not to the API spec.