Skip to content
MCP server

Pushary: human-in-the-loop for AI agents

Send your coding agent's approval requests and questions to your phone, Mac or Slack

Pricing
Free trial
Platforms
iOS, Android, macOS
Runs
Remotely
Version
1.5.2
Listing updated
2 days ago

About Pushary: human-in-the-loop for AI agents

Pushary: human-in-the-loop for AI agents is an MCP server for productivity. It runs as a hosted remote server, so there is nothing to install locally.

Pushary targets the moment an agent stops and waits for a yes. The agent asks through MCP, and the question reaches your phone, the Mac app's notch, Slack or a web app. You approve, deny or answer, and every answer is saved to an audit trail. It supports three question types: yes/no, multiple choice and free text. A detail page can carry file changes, error details and suggested next steps, and each notification says which agent is asking.

The rules are the distinctive part. Low-risk, reversible steps can be auto-approved, while deleting, deploying or spending money goes to you, and each step has a wait time after which your rule decides to deny or hand control back to the terminal. For Claude Code, permission hooks route tool approval prompts through push. The site says enforced approvals depend on the host integration, while questions and notifications work with any MCP client. It also says it sends the step that needs you rather than your codebase.

Setup is npx pushary@latest setup, which finds agents on the machine (Claude Code, Codex, Cursor, Gemini CLI, Hermes, Windsurf, VS Code, OpenCode and others), pairs a phone by QR code and configures them. The skill and adapters are MIT-licensed, while hosted delivery needs a paid plan. Partner plans let companies embed approvals for their own users.

Install

Run in a terminal

claude mcp add --transport http pushary https://pushary.com/api/mcp/mcp

Add to Settings → Connectors → Add custom connector

https://pushary.com/api/mcp/mcp
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "pushary": {
      "url": "https://pushary.com/api/mcp/mcp"
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "pushary": {
      "type": "http",
      "url": "https://pushary.com/api/mcp/mcp"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "pushary": {
      "serverUrl": "https://pushary.com/api/mcp/mcp"
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "pushary": {
      "type": "http",
      "url": "https://pushary.com/api/mcp/mcp"
    }
  }
}

What it does

Phone approvals
Approve, deny or answer an agent's question from iPhone or Android.
Risk rules
Auto-run safe steps and send irreversible ones to you.
Wait-time fallback
Choose whether an unanswered step is denied or handed back to the terminal.
Audit trail
Keep a record of who answered what across every agent.
Multiple channels
Phone, Mac notch, Slack and a web app.
One-command setup
npx pushary@latest setup connects the agents it finds.

Good for

  • Let a long coding task continue while you are away from your desk
  • Require a tap before a deploy or file deletion
  • Keep one rulebook across Claude Code, Codex and Cursor
  • Review which agent was approved to do what
  • Embed approvals in your own agent product for your users

Pricing

Free trial, from $9.99 a month

Agent plan $9.99/month after a 3-day free trial; higher plans and partner plans also offered.

Prices change. Check pushary.com before you buy.

Pushary: human-in-the-loop for AI agents FAQ

How much does it cost?

The Agent plan is $9.99 a month after a 3-day free trial; Agent Pro, Team, Partner and Enterprise plans are also listed.

Is my code sent to Pushary?

The site says Pushary sends the step that needs you, not your codebase.

Which agents work?

Claude Code, Codex, Cursor, Antigravity CLI, Gemini CLI, Hermes Agent, VS Code, OpenCode and other MCP agents.

How do I set it up?

Run npx pushary@latest setup, open the QR link on your phone and approve the connection.

Is any part open source?

The skill and adapters are MIT-licensed; hosted phone delivery requires a Pushary account.

Is Pushary: human-in-the-loop for AI agents free?

Pushary: human-in-the-loop for AI agents is a paid product with a free trial. Paid plans start at $9.99 a month.

How do I add Pushary: human-in-the-loop for AI agents to Claude Code?

Run this command in your terminal: claude mcp add --transport http pushary https://pushary.com/api/mcp/mcp

What are the best alternatives to Pushary: human-in-the-loop for AI agents?

The closest alternatives to Pushary: human-in-the-loop for AI agents are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to Pushary: human-in-the-loop for AI agents

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ha MCP

    MCP server

    MCP server for managing and automating Home Assistant from AI assistants

    Runs locally
    • 8 tools
    • Open source
    • 7.9K GitHub stars
    • Productivity