Elvean

A private Mac workspace where every AI model shares one desk

Intermediate
Screenshot of Elvean, A private Mac workspace where every AI model shares one desk

What is Elvean?

Elvean is an AI chat bot tool. A native Mac AI client that puts cloud and local models in one window, with projects that build a persistent knowledge base, live charts and maps in replies, threaded conversations, and a sandboxed Linux environment for agent tasks.

Elvean is a native macOS application that turns a chat window into a working environment. Instead of subscribing to one assistant, you connect the providers you already pay for, including OpenAI, Anthropic, Google, xAI, DeepSeek, Qwen and Cerebras, plus local runtimes such as Ollama, LM Studio and MLX. Through OpenRouter the app reaches more than 300 models, and any OpenAI compatible endpoint can be added. Keys are held in the macOS Keychain rather than on Elvean's servers, and conversations are not stored remotely. The centrepiece is the Project. Each project carries persistent system instructions, attached files and a filesystem style knowledge base that the model can search, read and write mid conversation. An auto-distill option extracts key facts into that base after each response, so context accumulates without a hand built retrieval pipeline. From there, agent mode chains tool calls: web and image search, Apple Calendar, Apple Notes, weather data and MCP connectors for services like Slack, Notion, Linear and Sentry. Answers arrive as sortable tables, Chart.js or TradingView charts, embedded Apple Maps and photo grids rather than plain prose. A sandboxed Linux container per project lets the model run commands, install packages and leave artifacts in the project folder. The fit is narrow but deep. Anyone already holding API keys, or running local models for privacy reasons, gets a single front end that replaces a stack of browser tabs. Threaded conversations, message forking, a prompt library and a system-wide hotkey panel suit people who research, analyse or code across long sessions. Free covers unlimited chat with any model. Pro at $4.99 monthly or $49.99 yearly unlocks Projects, search, rich rendering, MCP servers and exports. A $149 lifetime tier exists too. The obvious limits: it is Apple Silicon only, requires macOS 15 or later, has no Windows, web or mobile client, and can feel redundant for users who are content inside one vendor's app.

How do you use Elvean?

  1. 1Download and open the app
    Grab the disk image from the site, install it, and launch. No account, email or sign-up is involved, and a built-in demo key lets you send a first message immediately.
    Elvean — Download and open the app
  2. 2Add a provider key or a local model
    Open settings and either paste an API key from a provider such as OpenAI or Anthropic, or point the app at Ollama, LM Studio or MLX for offline use. The help guide covers the OpenRouter route, where one key unlocks 300+ models.
    Elvean — Add a provider key or a local model
  3. 3Create a project and seed the knowledge base
    Projects hold persistent system instructions and files. Add documents, PDFs or code to the project folder so the model can search and write knowledge base pages during the conversation.
  4. 4Turn on tools and connectors
    Enable web search, Apple Calendar, Notes and weather, then connect services like Slack through MCP connectors in settings. Agent mode will call these tools across multiple steps.
    Elvean — Turn on tools and connectors
  5. 5Read results as charts, tables and maps
    Pro responses include interactive Chart.js or TradingView charts, sortable tables and embedded Apple Maps. Edit any reply in markdown, fork a message, or branch a side thread to explore another direction.
  6. 6Run commands in the sandbox
    Open the terminal panel with Cmd+Shift+T to reach the project's isolated Linux container, where the model can install packages and leave artifacts in the project folder.
    Elvean — Run commands in the sandbox

Pros and cons

Pros

  • Connects cloud providers and local runtimes in a single interface, with mid conversation model switchingAI
  • Project knowledge bases that the model reads and writes, plus optional auto-distill of key factsAI
  • Responses render as sortable tables, interactive charts, Apple Maps and image galleries instead of plain textAI
  • Bring your own key model means paying providers directly with no markup, and keys stay in the macOS KeychainAI
  • Native SwiftUI build with a system-wide prompt hotkey, a prompt library and message forking for long research sessionsAI

Cons

  • Mac only and Apple Silicon only, requiring macOS 15 or later, with no Windows, web or mobile versionAI
  • Cloud models still need your own provider keys and separate billing, so total cost is not just the Elvean licenceAI
  • Visual rendering, web search, MCP servers and the Linux sandbox are gated behind Pro, so the free tier is chat onlyAI
  • Sandbox and agent modes assume comfort with API keys, MCP configuration and terminal conceptsAI
  • No team accounts, shared workspaces or admin controls are described on the siteAI

How much does Elvean cost?

Free trial

A 7-day free trial is offered on both the monthly and annual Pro plans, so the paid features can be tested before committing.

Learn more

Pricing

Free tier at $0 with unlimited conversations, 300+ cloud models via your own keys, local models, folders and the prompt panel. Pro costs $4.99 monthly or $49.99 yearly and adds Projects with knowledge base, search, rich rendering, MCP servers and exports. Lifetime is $149 once.

Learn more

Support

Pro subscribers get priority support, and the site publishes a help guide on obtaining API keys alongside a changelog and blog covering setup and workflows.

Learn more

Integrations

One-click MCP connectors for Slack, Notion, Linear, Sentry and Atlassian, plus MCP servers for files, databases and external APIs. Built-in Apple Calendar, Notes and WeatherKit tools, and search providers including Brave, Google, Serper, Exa, Parallel and Jina.

Learn more

Features

Projects with persistent instructions and a filesystem knowledge base, auto-distill of facts, 300+ models through OpenRouter plus local runtimes, agent mode with web search and Apple app tools, MCP connectors, a per-project Linux sandbox, threaded conversations and message forking, interactive charts, tables, maps and galleries, and a system-wide prompt panel.

Frequently asked questions about Elvean

  • How much does Elvean cost?
    Free tier at $0 with unlimited conversations, 300+ cloud models via your own keys, local models, folders and the prompt panel. Pro costs $4.99 monthly or $49.99 yearly and adds Projects with knowledge base, search, rich rendering, MCP servers and exports. Lifetime is $149 once.
  • Does Elvean offer a free trial?
    A 7-day free trial is offered on both the monthly and annual Pro plans, so the paid features can be tested before committing.
  • How do you use Elvean?
    The walkthrough on this page covers 6 steps: 1. Download and open the app 2. Add a provider key or a local model 3. Create a project and seed the knowledge base 4. Turn on tools and connectors 5. Read results as charts, tables and maps 6. Run commands in the sandbox.
  • What platforms does Elvean support?
    Elvean is available on MacOS.
  • What does Elvean integrate with?
    One-click MCP connectors for Slack, Notion, Linear, Sentry and Atlassian, plus MCP servers for files, databases and external APIs. Built-in Apple Calendar, Notes and WeatherKit tools, and search providers including Brave, Google, Serper, Exa, Parallel and Jina.
  • What are the limitations of Elvean?
    Mac only and Apple Silicon only, requiring macOS 15 or later, with no Windows, web or mobile version. Cloud models still need your own provider keys and separate billing, so total cost is not just the Elvean licence. Visual rendering, web search, MCP servers and the Linux sandbox are gated behind Pro, so the free tier is chat only.

Status

StatusActive
Views0
Outbound clicks0
Added10/22/2026

Platforms

MacOS

Pricing

FreemiumSubscriptionPaid