Migrating from Vertex AI?

Looking for an alternative to Vertex AI? Nyckel functions train in seconds (not hours), model deployment is free, and inference calls start at $0.005 per invoke.

Function Types

Input

Output

Examples

Input

What type of data do
you want to use?

Image

Text

Tabular

Output

What would you like to
do with your data?

Classify

Tag

Detect

OCR

Back

Text Classification

Train a function to categorize text data. Learn more >>

Text Tagging

Train a function to tag text. Learn more >>

Text Search

Create a gallery of text data. Search the gallery using images or text. Learn more >>

Image Classification

Train a function to categorize images. Learn more >>

Image Tagging

Train a function to tag images. Learn more >>

Image Search

Create a gallery of images. Search the gallery using images or text. Learn more >>

Image OCR

Extract text from images. Learn more >>

Object Detection

Train a function to detect objects in images. Learn more >>

Image Detection

Train a function to find the location and size of objects in an image using labels you provide. Learn more >>

Tabular Classification

Train a function to categorize tabular data. Learn more >>

Tabular Tagging

Train a function to tag tabular data. Learn more >>

Nothing for this one

Back

Select input and output to start creating a custom ML function based on your own data. Don't want to train your own function? Check out our pre-trained functions instead.

Stories From Some of Our Customers

Machine Learning Doesn't Have to Be That Hard

"Our AI moderator was live within a day and is now saving us $120k annually — a 20x return on investment."

Eyass Shakrah

Co-Founder, Pet Media Group

"It was awesome. After moving to Nyckel, we reduced the workload on our botanists by 70% even while growing our customer base by 2x."

Sunil Rawal

AI Lead, Gardyn

"Nyckel was so remarkably easy. My non-technical colleagues were able to jump in and quickly annotate the customer feedback to train the model."

Blake Kohler

CEO, Pulse for Good

"After moving to Nyckel, our average content moderation time went down 3x, auto moderation coverage increased 4x, and accuracy improved to 96%."

Vladislav Yavorskyi

Moderation Manager, Taimi

"Creating a custom spam filter with Nyckel was wonderful. We finally got rid of spam profiles that were draining our site ranking, bothering customers, and giving us a general headache."

Peter Swan

CEO, Gust

"Our Nyckel-powered 'magic' camera is a game highlight. And bringing it up was so easy. I went from creating a Nyckel account to a working prototype in literally ten minutes."

Gwyn Morfey

Tech Lead, SPYSCAPE

"AI-powered auto-moderation allowed quicker release of content to our public repository of credit card charges, driving a 4x increase in Ad revenue."

Jeremy Schneider

Co-founder, WTC

Why Nyckel?

Nyckel is optimized for non ML experts that want to ship ML features quickly.

Use our UI to quickly develop a POC, and then integrate via our API.

Nyckel

Vertex AI

Technology

Nyckel

Automatic re-train and re-deploy on new data.
Elastic deploy infra.

Vertex AI

Training and deploy only on user request.
Static deploy infra.

Pricing

Nyckel

Free train & deploy. Elastic pricing per invoke, starting at $50/month.

Vertex AI

Training ~$20 / model.
Deploy: >$1000 / model / month.

Time to POC

Nyckel

Minutes. Train first model in second. Review model predictions directly in UI.

Vertex AI

Hours. Wait 3-5 hours for first model to train. No ability to review predictions in UI.

Performance

Nyckel

Matching or exceeding VertexAI for images and text.

Vertex AI

On par with leading AutoML systems.

Nyckel in 60 Seconds

Train your own hosted Machine Learning model in just a few minutes. The model is instantly deployed and ready to be integrated in your application.

Explore Our API

Train your function through the UI, then call it through the API. Or do everything through the API.

Explore the API

Learn More on Our Blog

Image Classification Benchmark: Google, AWS, Hugging Face & Nyckel

We benchmark Nyckel against leading AutoML providers using several public datasets. We evaluate performance, training time, latency, throughput, and developer experience.

Why Narrow AI is Better Than GPT-4 For Machine Learning-Driven Decisions

GPT-4 is impressive. Yet, there are better tools available (including Nyckel) that are tailored to the unique needs of software developers and product teams who need an API for machine learning-driven decision making.

Computer Vision SaaS Landscape: Comparison of the Top 9 Players

In this article, we compare the functionality, amount of time it takes to set up a model, performance, positioning, and pricing of the nine leading computer vision SaaS players. Results show significant variations across accuracy, user experience, and pricing structures.

A top 3% score in the Kaggle Titanic Challenge using Transformers

In this post we show how to make a top 3% submission on the Kaggle Titanic Challenge using Transformers

Introducing Nyckel Image Search Functions

Nyckel Image Search functions enable cross-modal semantic search of images and text using the latest Deep Learning technology. Get started with a few simple API commands.

Custom AI-Assisted Content Moderation

Online communities need content moderation to thrive, but the diversity of the internet is not fully served by general purpose content moderation APIs. We lay out the need for custom AI-assisted content moderation, look at what such a moderation pipeline would look like, and show how Nyckel makes it significantly easier to build such a pipeline.

Whac-a-Spammer - How to Deal with Bots on Your Site

Bots posting content to your site can be a scourge. Stay one step ahead of them with these tips, tools, and tricks.

Building a Recommender system using the Nyckel API.

We break down the components of a Recommender system and show how ot set this up using the Nyckel API.

Pricing

Not sure if Nyckel is right for you? Give it a try today!

Open sign-ups
Always-free tier
First month free
Free model training

Development

Free

max 1,000 invokes / month

1,000 invokes included

Upgrade for more invokes

Up To

Functions: 10 /account

Classification

Samples: 5,000 /function

Labels: 20 /function

Search

Samples: 5,000 /function

Detection

Samples: 50 /function

Production

$50

min spend / month

10,000 invokes included

Additional invokes: $0.005

Up To

Functions: 1,000 /account

Classification

Samples: 10,000 /function

Labels: 50 /function

Search

Samples: 1,000,000 /function

Detection

Samples: 100 /function

Growth

$500

min spend / month

500,000 invokes included

Additional invokes: $0.001

Up To

Functions: 10,000 /account

Classification

Samples: 25,000 /function

Labels: 200 /function

Search

Samples: 10,000,000 /function

Detection

Samples: 500 /function

Enterprise

Custom

Scale to

Thousands of labels

Millions of samples

Billions of invokes

Plus

Single Sign-On

Deploy to VPC

Model export

Enterprise SLAs

Enterprise DPAs

Dedicated support

Invokes refer to invocation of trained functions using the /invoke, /search, or /locate endpoints. An additional charge of $0.0005 per Sample (Image, Text or Tabular data entry added to your function) per month is applied to each account.

Migrating from Vertex AI?

Function Types

Input

Output

Text Classification

Text Tagging

Text Search

Image Classification

Image Tagging

Image Search

Image OCR

Object Detection

Image Detection

Tabular Classification

Tabular Tagging

Machine Learning Doesn't Have to Be That Hard

Why Nyckel?

Nyckel in 60 Seconds

Explore Our API

Pricing

Development

Up To

Classification

Search

Detection

Production

Up To

Classification

Search

Detection

Growth

Up To

Classification

Search

Detection

Enterprise

Plus

Built By Industry Experts