Notes from the work

One email when a tutorial or migration write-up goes live. Nothing else, and an unsubscribe link in every email.

LeanZero

Two people in Romania doing Atlassian migrations, Forge apps and practical AI work for teams that would rather talk to the person doing the job. Most of what we learn ends up on this site.

Services

  • Atlassian migrations
  • Atlassian FastShift
  • Atlassian maintenance
  • Forge app development
  • AI development consultation

Topics

  • Jira
  • Jira Service Management
  • Confluence
  • Bitbucket
  • Atlassian Forge
  • Cloud Migration
  • Local AI
  • AI Coding
  • Certifications
  • Atlassian Team
  • All topics

Company

  • Blog
  • Tutorials
  • Contact

Community

  • Join Discord
  • Support this site

© 2026 LeanZero. All rights reserved.

Privacy policyTerms of service24/7 SupportTrust centerSecurityDPALegal notice
  1. Home
  2. Portfolio
  3. Elzi
  4. Features
  5. Providers
Elzi feature guide

Providers

Where models come from: the MLX engine on your Macs, or a cloud provider with your key.

Providers are where your models run: the LeanZero MLX engine on your own Macs, the cloud providers you've signed in to, and any OpenAI-compatible server you add. A provider is the source; to use one for work, you turn one of its models into a node. Elzi ships without models, so the first step is to download a local model or connect a cloud provider.

Local models with LeanZero MLX

The LeanZero MLX tab runs models on your own Macs with the MLX engine. It needs no credentials. The app walks you through three steps: connect your other Macs if you have them, get a model, and run it.

  1. Your Macs: connect your other Macs through LeanZero Link, or skip this to use only this Mac.
  2. Models: download a model from Hugging Face into the models folder. Each Mac has one models folder, used for downloads, copies and loads alike.
  3. Run it: start the model on this Mac, on another Mac or across two.

The tab also has Engine, Memory and Sampling settings. The engine serves one model at a time per Mac. LM Studio models already running on your network join the pool by themselves.

Cloud providers

The Cloud Providers tab lists the key-based providers Elzi supports: Claude (Anthropic), Amazon Bedrock, Azure Foundry, OpenAI, Gemini (Google), Qwen (Alibaba), OpenRouter, Ollama Cloud, MiniMax, Mistral AI, Z.AI, xAI, Moonshot and DeepSeek.

  1. Choose "Set up" next to the provider and paste your key.
  2. Elzi asks the provider for its model list. Pick the default model.
  3. Save. Saving runs the model once, which is how the key is proven.

Keys are encrypted into Elzi's secret store on your Mac. "Recheck connections" tests the saved providers again. Usage is billed by the provider, on your account.

OpenAI-compatible endpoints

"Add endpoint" connects any server that speaks the OpenAI API, such as vLLM, llama.cpp, Together, Groq or a company gateway. You can add as many as you need. These endpoints are chat only for now: chats and nodes can use them, swarm builds can't yet.

Related guides

Nodes and strategiesThe models Elzi hands work to, and the rules for which one does what.The LeanZero MLX engineThe local model server inside Elzi, built on our fork of Rapid-MLX.Serving a model on one MacElzi checks the memory, mounts the model and keeps long chats inside what fits.LeanZero LinkUse your other Macs: run a model there, copy models over, or split one across two.

Facts on this page were checked against our own test records on 6 November 2026.

Using Elzi

  • What Elzi is
  • Agent sessions
  • Software build sessions
  • Benchmark runs
  • Nodes and strategies
  • Providers

The MLX engine

  • The LeanZero MLX engine
  • Serving a model on one Mac
  • Distributed serving across two Macs
  • N-gram tables and prompt lookup
  • Speed: measured numbers
  • Memory lock

Media and devices

  • Images
  • Voices
  • LeanZero Link

Elzi

  • All feature guides
  • Download and user manual