Glint AI Studio

Based on Glint Agent, build autonomously evolving visual intelligence solutions

What visual algorithm project can we help with?

Recommended

Fromone requesttoa deliverable algorithm

01

Understand

  • Describe the scene in natural language
  • Agent returns requirement understanding
  • Recommend approach and path

02

Create project

  • Prefill project type
  • Name and label categories
  • Spin up a standard workspace

03

Prepare data

  • Upload, capture, and reflow data
  • Generate and centralize samples
  • Govern quality and analysis

04

Annotate & train

  • Manual and AI annotation
  • Dataset versioning
  • Multi-path model training

05

Convert & deploy

  • Finish model testing
  • Complete format conversion
  • One-click edge deployment

06

Reflow & iterate

  • Continuous on-site sample reflow
  • Consolidate into the data center
  • Drive ongoing algorithm evolution

Product videos

Analysis

03:42

Start an algorithm project with one sentence

Glint Agent understands the scene, recommends a plan, and creates executable project settings.

Data

02:18

Bring multi-source data into the training line

Delivery

03:07

From data version to one-click delivery

Optimize

02:45

Let field data drive continuous evolution

Product services

01

Cloud training

Cloud training pack

Cloud training pack
TermTraining hoursObject storageValidity
Monthly50 hours100 GB1 month
Half-year300 hours600 GB6 months
Annual600 hours1.2 TB12 months
EgressMatches the pack's object storage

Included

  • Cloud training workspace
  • Train, evaluate, and version models
  • Store data and training artifacts

Optional add-ons

02

Train + deliver

Cloud training and release

Full cloud delivery
Cloud training and release
TermTraining hoursObject storageValidity
Monthly50 hours100 GB1 month
Half-year300 hours600 GB6 months
Annual600 hours1.2 TB12 months
Release licensePer algorithm / project for target devices

Included

  • Everything in the cloud training pack
  • Convert, publish, and manage algorithm packages
  • Target-device licenses and delivery support

Optional add-ons

03

On-prem delivery

Private deployment

Delivery
Project implementation and acceptance
Environment
On-prem / private cloud
Platform
Data, training, evaluation, and model management
Data boundary
Data and models stay in the customer environment

Included

  • On-prem platform software deployment
  • Appliance or customer-server adaptation
  • Install, training, and acceptance support

Optional add-ons