AION
ProductPricingSkillsDocsAboutSecurity
Sign inStart free
Sign inStart free
search
menu_bookDocumentation
Basics2
  • rocket_launchGetting started
  • workspacesSpaces and permissions
Work3
  • task_altTasks
  • forumDiscussions
  • folder_openDocuments and knowledge
AI3
  • smart_toyAgents
  • hubModels: all in one place
  • micVoice and AI chat
Connections2
  • extensionMCP and integrations
  • boltAutomations
Platform3
  • devicesApps
  • shieldSecurity and data
  • helpQuestions and answers
  1. Documentation
  2. AI
  3. Models: all in one place
hub

Models: all in one place

Choose a model for the task, not for the vendor. Every agent can run on its own.

On this page
  • How it works
  • Models
  • Why it helps
  • Get started
  • Where to find it
  • Details
Learn by doing

A practice workspace with step-by-step guidance and result checks. No sign-up or account changes.

How it works

AION accepts keys from six providers: OpenAI, Anthropic, Google, xAI, OpenRouter and any server compatible with the OpenAI API.

A key is personal or shared with a project. It is stored encrypted on the server, and only the last 4 characters ever reach the browser.

OpenRouter lets you choose models from its available catalogue. Availability depends on the provider and connection. OAuth sign-in is supported without manually copying a key.

Ollama and LM Studio run generation locally through AION for Mac, without automatic cloud fallback. Workspace tasks, documents and collaboration still use the AION server.

With a provider key
  • GPT
  • Claude
  • Gemini
  • Grok
Through OpenRouter
  • OpenRouter
  • DeepSeek
  • Llama
  • Mistral
  • Qwen
  • Kimi
  • GLM
Locally on a Mac
  • Ollama
  • LM Studio

Why it helps

  • checkChoose an available model for the task and check its cost with the provider.
  • checkYour own key does not use your plan quota.
  • check“Check connection” tells you at once if the key is wrong or the model is not found.

Get started

  1. 1

    Open “AI models” or Space → Settings → Assistant settings.

  2. 2

    Add a provider key — personal or for the project.

  3. 3

    Click “Check connection”.

  4. 4

    In the agent's “Model” field, choose “Choose an API model” and set the provider and model, or keep “Use the workspace default”.

  5. 5

    Optionally add “Spare routes” in case the main model is unavailable.

Where to find it

  • subdirectory_arrow_rightIntegrations → Your models
  • subdirectory_arrow_rightAI models
  • subdirectory_arrow_rightSpace → Settings → Assistant settings
  • subdirectory_arrow_rightAgent → Model

Details

Without your own key

  • Requests go through the default model and use your plan quota: one unit is about 1,000 tokens.
  • Default quotas are Free: 100 units, Pro: 2,000, Team: 6,000. Server configuration and the active plan may change limits; check Billing.
arrow_backPreviousAgentsNextarrow_forwardVoice and AI chat
Didn't find an answer?

Write to support and we will reply by email. Or just try it: your personal space is ready right after sign-up.

Start for freeSupport

Model marks: lobehub/lobe-icons, MIT licence. Names and logos belong to their owners and are shown only to name the models AION works with.

Checked against the code on 4 November 2026

AION

A workspace for people and AI agents.

PRODUCTFeaturesPricingSkillsDocs
COMPANYAboutSecuritySupport
LEGALPrivacyTerms of use
© 2026 AION · Built for teams that work with agents