The Production AI Platform
Everything you need to build, test and deploy Production AI
Store and search vectors and inputs. Auto-label data. Fine-tune models. Deploy and iterate.
Clarifai is a unified, full lifecycle platform accelerating enterprise-grade production AI for the most demanding tasks.
- Overview
- AI Lake ™
- Data Preparation
- Model Development
- Model Inference and Ops
- AI Lake™ is the reference for all your AI assets.
- Spacetime lets you search vectors and inputs.
- Enlight to fine-tune any open source or third-party model.
- Armada immediately deploys your models at production scale.
- Extend deploys your Streamlit UIs on top of Clarifai's Portal.
- AI Lake™: Organize, share, and reuse all your inputs, vector embeddings, datasets, annotations, models, workflows, and modules. One central place for teams to collaborate.
- Spacetime Vectors and Search: Create, store, and search vector embeddings and inputs.
- Scribe Automated Data Labeling: Automation-first labeling to label data faster and more accurately.
- Enlight Training and Evaluation: Easy, customizable UI for training and evaluating models and managing model versions over time.
- Armada: Auto-scaling model inference engine for production.
- Mesh: Drag-and-drop workflows engine.
- Extend: Streamlit UIs for infinite tasks.
- Collectors: Production data collection for continuous learning.
Open, reusable building blocks for the full AI lifecycle
Clarifai’s fully integrated AI platform enables a diverse portfolio of machine learning products and services across industries that work seamlessly together to create AI at scale.
Data Labeling
Automate data labeling with Scribe
Automation-first flows for high-quality training data.
- Auto-label as you ingest
- Scale to labeling projects of any size
- Labeling and reviewer UIs maximize quality and quantity of annotations
Data Management
Manage and search data with Spacetime
Patented search and lifecycle management for inputs and vectors.
- Vector similarity, keyword, and metadata search
- Rank, filter, and sort your data by human annotations, AI predictions, metadata, and geo-location
- Transform searches into datasets
Model Training
Train & evaluate models with Enlight
Train, evaluate, and manage AI models and model versions at scale.
- Use prebuilt models from Clarifai, our community, or upload your own third party models
- Train new models with only a few examples in seconds
- Maximize model performance by training your own custom embedding
MLOps
24/7 AI without DevOps with Armada
Scale compute to run your models at production quality cost-effectively.
- Adds or removes replicas based on the number of messages and the capacity for each server instance
- Optimized runtime to maximize GPU utilization of each model replica
- Monitors in-flight requests to each model within the platform for timely processing
Model Workflow Engine
Easily process complex logic with Mesh
Drag-and-drop interfaces tie models and logical operators together and automate complex workflows.
- Express business logic as a computation graph of building blocks that runs as a single API call
- Turn predictions made by your AI models into action triggers
- Get highly optimized parallel processing of data and workflow nodes with no DevOps
Streamlit UI Modules
Build custom UI modules with Extend
Streamlit UI modules for popular ML tasks.
- Collaborate on reusable UI components across your organization or the broader AI community
- Build UIs for colleagues to perform valuable tasks
Collectors
Automated data collection
Collect real-world prediction data from models using custom criteria.
- Isolate data for analysis, labeling, and evaluation
- Power continuous learning loops
- Scribe Label
- Spacetime Search
- Enlight Train
- Armada Predict
- Mesh Workflows
- Extend
- Collectors
Centralized and secure AI
Utilize foundation models to streamline the process of developing and deploying machine learning models. Efficiently organize, share, and reuse all your AI resources across your enterprise.
AI Lake™
Accelerate AI adoption and reusability across the enterprise
Enable teams to collaborate more easily and efficiently while providing access control of all your AI models, annotations, datasets, workflows, and UIs.
- Store and analyze massive volumes of data objects
- Automatically index your data with AI
- Track lineage and versioning of all your AI assets
Flexible deployment options
Clarifai offers any combination of deployment options. The platform can be deployed on any cloud (AWS, Azure and Google Cloud), on bare-metal with air-gapped options, and on the edge.
Multi-cloud
Bare-metal
IoT + Mobile
Deploy AI to edge devices with Flare
Take advantage of the complete Clarifai platform plus additional tools and features designed for edge deployments.
Operationalize a large number of custom AI models
Document question answering
Answer questions based on customizable contexts.
Semantic search
Use AI to search inside your unstructured data.
Metadata generation
Auto generate metadata from your unstructured data.
Face recognition
Use faces to identity people, authenticate, detect deepfakes, gauge emotion or demographics.
Content moderation
Take the human out of the content moderation loop.
Video security
Generous maternity, paternity and adoption leaves, as well as family medical care leave.
Visual inspection
Receive a monthly stipend towards your cell phone bill.