OneOpen ML Studio Logo

Product

  • Product vision
    • Definition
    • Who it is for
    • Platform scope
    • Modalities
    • What we are not
  • Principles
    • Local-first
    • Self-hosted by default
    • Framework-neutral
    • Modality-independent architecture
    • Reproducible
    • Extensible
  • Architecture
    • High-level platform
    • Platform modes
    • Organizational hierarchy
    • Universal dataset schema (direction)
    • Plugin surfaces (target)
    • Training adapters
    • Suggested modular backend layout (evolution)
    • Technology direction
  • Modules
    • Annotation modalities
      • Available today (visual / CV)
      • Other modalities today
    • Roles
  • UI design system
    • Tokens
    • Shell
    • Components
    • Annotate (visual)
    • Pipelines
    • Out of scope (UI)
  • Roadmap
    • Shipped
      • Computer vision
      • Text, tabular, audio, video, LLM
      • Collaboration & ops
    • Planned
    • Guides
  • Package rename notes

Getting started

  • Installation
    • From PyPI (recommended)
      • Optional extras
    • From source
    • Verify
    • Platform notes
    • Build this documentation locally
  • Quick start
    • 1. Install
    • 2. Initialize data directory
    • 3. Start the server
      • Useful options
    • 4. Create a project
    • 5. Add classes and images
    • 6. Annotate (visual)
    • 7. Version, train, or export
    • Typical first workflow (CV)
    • Next steps
  • Concepts
    • Platform hierarchy
    • Today
    • Projects
    • Classes, images, annotations (visual)
    • Splits and versions
  • Deploy with Docker
    • Quick start
    • Services
    • Health
    • Local development without Docker
  • Active learning and enterprise
    • Active learning
      • Methods
    • Kubernetes
    • First-time setup
    • Enterprise stubs
    • Non-goals
  • Pipelines and plugins
    • Plugin registry
    • Pipeline schema
    • CRUD and YAML
    • Running a pipeline
    • UI
  • Text and tabular modalities
    • Project types
    • Import
    • Samples API
    • Export
    • Training adapters
  • Audio, video, and LLM datasets
    • Modalities
    • Content shapes (JSONL)
    • APIs
    • UI
    • Out of scope (for now)

User guide

  • Projects
    • Create a project
    • Open and update
    • Delete
    • Classes
    • Dashboard stats
  • Annotation
    • Computer vision (visual)
      • Prerequisites
      • Project types
      • Tools
      • Workflow tips
      • Annotation sources
      • CV annotation API
      • Upload images
      • Delete images
    • Non-image modalities
  • Data cleaning
    • Where it fits
    • UI
    • API
    • Pipelines
    • Training gate
  • Deployment modes & plugins
    • First-time setup (Team & Enterprise)
    • Organizations & workspaces
    • Plugin manager
  • Data prep & training by task (with screenshots)
    • Studio overview
    • Computer vision
      • Object detection
      • Image classification
      • Instance & semantic segmentation
    • Text & tabular
      • Text classification
      • Named entity recognition (NER)
      • Tabular classification
      • Tabular regression
    • Audio & video
      • Audio classification
      • Transcription (ASR)
      • Video classification
      • Action recognition
    • LLM datasets
      • Instruction tuning (SFT)
      • Preference (DPO / ORPO)
      • RAG evaluation
    • Quick reference
    • Related docs
  • Label Assist (YOLO)
    • Requirements
    • How to run (UI)
    • Behavior details
    • API
    • Choosing a model
  • SAM (interactive segmentation)
    • Install
    • Checkpoint
    • Usage
    • API
    • Troubleshooting
  • Import and export
    • Import an existing dataset
      • Expected layouts
      • UI
      • API
    • Assign splits
    • Export
      • YOLO (zip)
      • COCO (JSON)
  • Dataset versions and pipeline
    • Create a version
    • Pipeline defaults
      • Preprocessing (typical defaults)
      • Augmentation
    • When to create a new version
    • Relation to training and export
  • Training
    • Requirements
    • Training by modality
    • Plugin manager (auto-install)
    • Before you train
    • Start a job (UI)
    • API
    • Device values
    • Windows note
    • Limitations
    • Troubleshooting
  • Models
    • Catalog
    • Register custom weights
    • Training runs
    • Project-scoped listing

Reference

  • CLI reference
    • Commands
      • oneopen start (recommended)
      • oneopen serve
      • oneopen init
      • oneopen version
      • oneopen user create-admin
      • oneopen worker
    • Future CLI (roadmap)
  • Configuration
    • Core
    • Deployment & auth
    • Database, queue, storage
    • Plugins & enterprise stubs
      • Examples
    • Derived paths
    • Binding beyond localhost
  • Data layout
    • Directory tree
    • SQLite tables (overview)
    • Backup
    • Disk use
  • HTTP API
    • Pages (HTML)
    • Projects
      • ProjectCreate
    • Classes
    • Images
    • Annotations
      • AnnotationCreate
    • Auto-annotate
    • Splits, versions, pipeline defaults
    • Plugins & pipelines
    • Text & tabular samples
    • Export
    • Models & training
      • TrainRequest
    • Identity & collaboration
    • Setup (Team / Enterprise)
    • Platform
    • Stats
    • Example: create → upload → annotate (curl)
  • Troubleshooting
    • Server will not start
    • UI errors / blank pages
    • Label Assist fails
    • SAM fails
    • Training fails
    • Import issues
    • Export issues
    • Where to get help

Project

  • License and credits
    • Citation
    • Third-party
    • Author
  • Changelog
    • 1.0.0 — first PyPI release
      • Platform
      • Hierarchy & Team / Enterprise
      • Security (Team / Enterprise)
      • Docs
    • 0.9.0
    • 0.8.0
    • 0.7.0
    • 0.6.0
    • 0.5.0
    • 0.4.0
    • 0.3.0
    • 0.2.0
    • 0.1.1
    • 0.1.0
OneOpen ML Studio
  • Search

© Copyright 2026, oneopensource.

Built with Sphinx using a theme provided by Read the Docs.