Build AI features, remarkably fast
Compose state of the art AI models into production ready APIs
Skip the papers, GPUs, python, servers, etc.
Over 100+ Models available in our library.
Audio and Text
Get a powerful toolbox for Sentiment Analysis, User Intent, Emotion Detection, and Entity Detection, by easily combining Audio and NLP models together.
Video and Images
Unlock metadata from your rich media to power Moderation, Search & Indexing, Labeling, with state of the art Transformer and Visual Question-Answer models.
LLMs and Generative AI
Safely tap into the latest models with privacy features such as Personal Information Masking, API Audit Logging, and other accountability tools for teams.
OpenAI ChatGPT
The Classic ChatGPT API from OpenAI for chat completion. Use our API or bring your own key.
OpenAI GPT3 Family
The Classic GPT3 model family for text completion. Use our API or bring your own key.
Anthropic Claude-100k
The latest Claude-100k model from Anthropic is great for huge document of up to 100k tokens. Access is case by case, or bring your own key.
Vicuna-13B
Vicuna is a 13B param LLM that is "90% of the way" to ChatGPT. Its small, fast, and license allows commercial use.
Detect Sentiment in Text
Given a body of text, automatically classify between various sentiment labels with confidence scores.
Extract User Intent from Text
Given a body of text, automatically extract the most likely user intents with confidence values.
Answer a Question from Text Context
Given a body of text, provide a text answer using the text as relevant context.
Detect People or Locations in Text
Given a body of text, automatically detect any references to Named Entities such as people, locations, and organizations.
Remove Stop Words from Text
Given a body of text, automatically remove unimportant stop words such as "and", "a", "the", "then", etc.
OpenAI ChatGPT
The Classic ChatGPT API from OpenAI for chat completion. Use our API or bring your own key.
OpenAI GPT3 Family
The Classic GPT3 model family for text completion. Use our API or bring your own key.
Anthropic Claude-100k
The latest Claude-100k model from Anthropic is great for huge document of up to 100k tokens. Access is case by case, or bring your own key.
Vicuna-13B
Vicuna is a 13B param LLM that is "90% of the way" to ChatGPT. Its small, fast, and license allows commercial use.
Detect Sentiment in Text
Given a body of text, automatically classify between various sentiment labels with confidence scores.
Extract User Intent from Text
Given a body of text, automatically extract the most likely user intents with confidence values.
Answer a Question from Text Context
Given a body of text, provide a text answer using the text as relevant context.
Detect People or Locations in Text
Given a body of text, automatically detect any references to Named Entities such as people, locations, and organizations.
Image to Labels with Confidence Scores
Given an image and a set of labels, automatically calculate the probability that a label is correct.
Image to Aesthetic Score
Given an image, automatically estimate how beautiful an image is based on 200k human labeled images from LAION.
Image to NSFW Score
Given an image, automatically detect adult content in images based on 200k human labeled images from LAION.
Image to Feature Embedding
Given an image, calculate an embedding / vector that captures the core features of this image. This embedding can be stored in a vector database and used for search.
Answer Questions Based on an Image
Given an image and a text question, provide a text answer using the image as relevant context.
Face to Emotion mapping.
Given an image with a human face, calculate the probability that certain emotions are present such as Happy, Sad, Angry, etc.
Speech to Text
Given an audio stream or audio file of any format, automatically transcribe the contents into text.
Text to Speech
Given a body of text, generate an audio file of a synthetic voice reading the text.
Detect Speaker Changes
Given an audio file with up to 3 different speakers, automatically detect different speakers and the times that they are speaking.
Image to Labels with Confidence Scores
Given an image and a set of labels, automatically calculate the probability that a label is correct.
Image to Aesthetic Score
Given an image, automatically estimate how beautiful an image is based on 200k human labeled images from LAION.
Image to NSFW Score
Given an image, automatically detect adult content in images based on 200k human labeled images from LAION.
Image to Feature Embedding
Given an image, calculate an embedding / vector that captures the core features of this image. This embedding can be stored in a vector database and used for search.
Answer Questions Based on an Image
Given an image and a text question, provide a text answer using the image as relevant context.
Face to Emotion mapping.
Given an image with a human face, calculate the probability that certain emotions are present such as Happy, Sad, Angry, etc.
Speech to Text
Given an audio stream or audio file of any format, automatically transcribe the contents into text.
Text to Speech
Given a body of text, generate an audio file of a synthetic voice reading the text.
Compose production ready workflows using our simple node editor.
A fully hackable editor means Jumbo won't limit what you can do. Define custom functions for data transformations, API requests, webhooks, and much more.
Production ready is only a click away with Jumbo's managed infrastructure. We'll spin up and down resources AWS and GCP to match your workload demands while optimizing costs.
Simple compute based pricing for organizations of all sizes
No usage? No bills. Pay only for the compute you use.
Development
Everything you need to build, test, and even deploy your next AI powered feature.
Starts at
$24/ month / seat
Small GPU
4x cores + 16GB RAM + 24GB GPU
+ $0.04 per minute of runtime
Large GPU
8x cores + 64GB RAM + 40GB GPU
+ $0.16 per minute of runtime
Performance GPU
Custom + 80GB H100 GPUs
+ Custom Pricing
Frequently Asked Questions
What cloud providers does JUMBO support?
JUMBO is platform agnostic and is designed to be deployed on any cloud provider including but not limited to AWS, GCP, and Azure. If your company has a specific cloud provider requirement, please contact sales for details.
I don't see my use-case covered, can JUMBO still help?
Absolutely! We have worked with clientel from major technology VCs such as Y-Combinator, Andressen Horowitz, to name a few. Reach out to mike@jumbo.io and we would be happy to dicuss your use-case and potential solutions.
Are custom models supported by JUMBO?
Yes there are multiple ways to use your custom models or model weights with JUMBO. You can upload custom weights to our platform if the model already exists, we can add support for your specific model, or you can use JUMBO through an on premise deployment and have hands on control of the entire platform in your own cloud environment.