Skip to content
See availability
Docs
  • Get started
  • Models
  • Gateway
  • Security
  • What is Ethen
  • Quickstart
  • Core concepts
  • Models overview
  • Model Library overview
  • Model Intelligence profiles
  • Choosing a model
  • Local versus hosted models
  • Model Library
DocumentationBrowse documentation+
  • What is Ethen
  • Quickstart
  • Core concepts
  • Models overview
  • Model Library overview
  • Model Intelligence profiles
  • Choosing a model
  • Local versus hosted models
  • Model Library
  1. Documentation
  2. Local Models

Local Models Docs

Understand where local model work fits.

Local Models Docs explain local lanes, open-weight models, runtime requirements, and how local work can fit alongside Gateway, Code, and workflows. Local execution depends on the device, runtime, model availability, resources, and workspace configuration.

See availabilityExplore Ethen Local
On this page+
  1. Local work depends on the setup.
  2. Core concepts
  3. How local model work fits
  4. FAQ

Local work depends on the setup.

A local lane can keep some model execution closer to the user where supported. It does not automatically make the entire Ethen experience offline, and it is not the right path for every task.

  1. 01

    Local lanes can support private or controlled work where the runtime makes them available.

  2. 02

    Open-weight models can support experimentation and controlled workflows.

  3. 03

    Behavior depends on device, runtime, model files, available resources, and workspace settings.

  4. 04

    Local and cloud paths can complement each other through Gateway and workflow design where supported.

Core local model concepts

Use these concepts to decide whether a local path fits the work and what must be configured first.

Local lanes
Use a local lane for sensitive or controlled context where local support and runtime configuration are available.
Open-weight models
Open-weight models provide another option for experimentation, controlled workflows, and cost-aware iteration when the configured runtime supports them.
Runtime requirements
Local behavior depends on the device, runtime, model files, resource limits, model availability, and workspace configuration.
Hybrid workflows
Local steps can fit alongside flagship or open lanes, Gateway routing, and repeatable workflows where supported. Local outputs can still feed into broader reviewable workflows when the surrounding product surfaces support that path.

How local model work fits

  1. 01

    Identify the context

    Decide whether the task involves private code, internal notes, customer-related context, or draft material that may benefit from a local path.

  2. 02

    Check local support

    Confirm the device, runtime, model availability, resources, and workspace configuration for the intended work.

  3. 03

    Choose the lane

    Use a local lane where supported, or choose flagship or open lanes when they better fit the task.

  4. 04

    Review the result

    Keep the model choice, relevant context, outputs, and review notes connected to workspace history where available.

  5. 05

    Connect the workflow

    Use Gateway and Workflow concepts when local model steps need to become part of a repeatable process.

  6. 06

    Scope

    Local Models Docs explain local lanes, open-weight models, runtime requirements, and workflow fit. Exact setup and compatibility depend on the current product, model, device, and runtime configuration.

FAQ

They explain local lanes, open-weight models, runtime configuration, limitations, and how local work fits into Ethen. They are the docs for understanding local lanes, open-weight model workflows, runtime assumptions, and private-context fit inside Ethen.

Not always. Local lanes and full offline behavior are different. Exact behavior depends on runtime configuration and product support. Because users need a grounded view of local tradeoffs before they rely on the lane.

Open-weight models are models whose weights are available for direct use under their applicable licenses. Only where the runtime and product surface support that behavior. The docs avoid stronger claims than the implementation proves.

Use a local lane where supported for sensitive context, private review, controlled experimentation, or workflows that benefit from staying closer to the environment. Yes. Many teams will use both, depending on the task and the sensitivity of the context.

Local lanes can fit into Gateway routing where the runtime and route configuration support them. Runtime support, context boundaries, lane fit, and review posture.

Model compatibility depends on the selected model, runtime, device resources, and workspace configuration. They help users choose local work intentionally instead of treating it like a vague privacy label.

Understand the local path before you depend on it.

Use Local Models Docs to understand local lanes, runtime requirements, open-weight models, and where local work fits inside Ethen.

See availabilityExplore Ethen Local

On this page

  1. Local work depends on the setup.
  2. Core concepts
  3. How local model work fits
  4. FAQ

Ethen is UpCube's family of independent AI products, sharing one governed Platform.

Products

  • Overview
  • Chat
  • Code
  • Studio
  • Research
  • Founder
  • Designer
  • Voice
  • Local Models
  • Gateway

Platform

  • Overview
  • Flow
  • Sentinel
  • GPU Compute
  • Computer
  • Approvals & Evidence
  • Security

Research & Models

  • UpCube Research Lab
  • Publications
  • Models guide
  • Model Library
  • Model Intelligence
  • Faros

Resources

  • Resources hub
  • Blog
  • Docs
  • Changelog
  • FAQ
  • Help

Company & Trust

  • About
  • Engineering
  • Transparency
  • Availability
  • Careers
  • Contact
  • Solutions
  • Pricing

Legal

  • Privacy
  • Terms
  • Acceptable Use
  • Cookies
  • Security
  • Third-party notices

© 2026 UpCube

  • Instagram
  • X
  • TikTok
  • YouTube
  • Facebook
  • LinkedIn

Status registry dated 2026-10-31