OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Workflow Agents

Mercury Agent

Mercury Agent is a self-hostable, persistent AI assistant written in TypeScript. It combines built-in tools, extensible skills, permission controls, token budgets and SQLite-backed memory, and can be accessed through a CLI and configured messaging or web channels. Its documented features also include Kanban workflows, scheduled tasks, a repository-aware coding environment and specialist bots.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Mercury Agent — OpenViST project profile in Workflow Agents.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Mercury when you want an assistant that can remain available across sessions and channels, retain structured memory, and perform tool-based work under configurable permissions. Its coding environment and specialist bots are relevant when tasks call for repository-aware work or delegation.

Who it is for

  • People who want a persistent assistant reachable from a CLI or supported messaging channel.
  • Users who need tool-using workflows with configurable approvals, role scopes and token budgets.
  • Developers seeking an agent-based coding environment with planning and delegated background agents.
  • Teams or individuals who want specialist bots with separate personas, scoped memory and permissions.

Documented uses

  • Send tasks to an assistant through CLI, Telegram, Discord, Slack or Signal.
  • Use built-in tools and installed skills while retaining control through approval modes and permission scopes.
  • Keep structured long-term context in SQLite-backed memory.
  • Organize work with Kanban boards and scheduled tasks.
  • Plan and execute repository work in Mercury Code, or delegate tasks to persistent specialist bots.
Before you choose

Requirements & limits

The project is open source under the MIT license. The README documents standalone installers for macOS, Linux and Windows, plus npm installation requiring Node.js 22 or later.

  • Choose a model provider and configure its credentials during setup.
  • Decide whether to use approval prompts or the session-level “Allow All” mode.
  • Messaging channels have distinct setup and access models; Signal uses a signal-cli bridge and the README notes platform-specific constraints.
  • The npm installation path requires Node.js 22 or later; standalone installers are also documented.
  • Mercury Code is a separate full-screen coding environment within the agent, while Mercury Bots are persistent specialist agents.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install and complete first-run setup

Install Mercury Agent and configure its initial name, model provider and permission mode.

Before you start

  • Choose a standalone installer for macOS, Linux or Windows, or use npm with Node.js 22 or later.
  • Have credentials for the model provider you intend to configure.

Steps

  1. For macOS or Linux, run the documented installer command: the command below.

    Documented command
    curl -fsSL https://mercuryagent.sh/install.sh | sh

  2. For Windows, run the documented PowerShell installer: irm https://mercuryagent.sh/install.ps1 | iex.

  3. Alternatively, with Node.js 22 or later, start with the command below, or install globally with the command below and run mercury.

    Documented command
    npx @cosmicstack/mercury-agent
    Documented command
    npm i -g @cosmicstack/mercury-agent

  4. Complete the first-run wizard with a name, provider and any optional channels, then select Ask Me or Allow All before chat starts.

Expected outcome

Mercury opens its CLI chat after initial configuration and a permission mode has been selected.

Guide limitations

  • The standalone installer commands are documented for macOS/Linux and Windows; the npm route requires Node.js 22 or later.
  • This guide covers initial configuration, not connecting and testing each messaging channel.
Documentation 1 ↗
Source-based practical guide

Reconfigure Mercury from the CLI

Open Mercury’s documented configuration and platform diagnostic commands after initial setup.

Before you start

  • Mercury Agent must already be installed and available as the mercury command.

Steps

  1. Run mercury doctor to reconfigure settings such as provider keys, name or other settings.

  2. Run mercury doctor --platform to request the platform-specific diagnostic.

Expected outcome

Mercury opens its documented reconfiguration or platform diagnostic flow.

Guide limitations

  • The documentation lists these commands but does not specify the diagnostic checks or their results.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
10 October 2026
Original website title
cosmicstack-labs/mercury-agent · Official homepage title · retained in its source language
How we review sourcesSuggest a correction