Honest comparison
Remote Control is Anthropic's own way to drive a Claude Code session on your computer from the Claude app. It comes with Pro, Max, Team and Enterprise plans. Here is how to start it, what it needs, and what to use when it is not open to you.
If your work is a single Claude Code session and you want the deepest, first-party control surface, Anthropic's Remote Control is excellent. If you run more than Claude, or you need one set of rules and one record across agents, Pushary is the better fit.
A first-party tool covers its own agent. Remote Control is for Claude Code. Pushary hooks into Claude Code, Codex, Cursor and Gemini CLI with one policy, and Hermes and other MCP clients can ask you through the same app. One config, one approval queue, one audit trail, no matter which agent is blocked. Vendors structurally will not build that for each other.
Claude Code has its own allow, ask and deny rules, and Remote Control can push when it needs you. Pushary lets you write the rules once for every hooked agent: auto-approve safe reads, push on Bash, escalate git push, always deny rm -rf. The phone only buzzes for the decisions that matter, and the rule holds across sessions, machines and agents.
Pushary keeps an immutable, attributed log of every question and every human decision across your agents and machines, exportable on Agent Pro and above. That is what a team lead or a compliance owner needs, and it is the kind of cross-vendor record a single-vendor remote does not set out to be.
First-party wins on depth and support for its own tool. If you only run Claude Code and want the most native control surface, backed by Anthropic, Remote Control is the right pick. Pushary makes the opposite bet: be vendor-neutral, and add the policy and audit layer that a single-vendor tool will not build for its competitors.
| Feature | Pushary | Remote Control |
|---|---|---|
| Works across vendors (Cursor, Codex, Hermes, any MCP client)CROSS-VENDOR | ||
| Works on API keys, Bedrock, Google Cloud and FoundryANY AUTH | ||
| One per-tool policy for Claude Code, Codex, Cursor and Gemini CLIPOLICY | ||
| Immutable audit trail across agents (export on Agent Pro and above)AUDIT | ||
| Works with Claude Code | ||
| Approve from your phone | ||
| Slack approvals (channel or DM) | ||
| Stop a running task from your phone | ||
| Stop every hook-connected agent at once (Claude Code, Codex, Cursor, Gemini CLI) | ||
| First-party, built and supported by Anthropic | ||
| Deepest possible Claude Code integration |
Run claude --remote-control (or claude --rc) to start a session you can also drive from your phone, or type /remote-control in a session that is already open. Then scan the QR code it shows, or find the session in the Code tab of the Claude app. claude remote-control on its own starts server mode, which can run several sessions at once.
No. Anthropic's own requirements list it as available on Pro, Max, Team and Enterprise plans and state that API keys are not supported, and it is unavailable when you use Amazon Bedrock, Google Cloud's Agent Platform, or Microsoft Foundry. On Team and Enterprise it is also off until an Owner enables the Remote Control toggle. This is the one gap that has nothing to do with features: if your Claude Code does not sign in through claude.ai, Remote Control is not an option at all. On a Console API key, Anthropic's Channels research preview lets you chat with the session from Telegram, Discord or iMessage; on Bedrock, Google Cloud and Foundry there is no first-party option. Pushary's hook runs locally on the tool call, so it does not care how Claude Code authenticates or which cloud serves the model.
No. It covers Pro, Max, Team and Enterprise, so any comparison still describing it as a Max perk is out of date. For most people the questions that decide it now are whether you sign in through claude.ai at all, and whether an Owner has enabled the toggle on Team and Enterprise. Gateways, Zero Data Retention and the HIPAA configuration also rule it out, and a long-lived setup token cannot start it.
It is Anthropic's own, first-party way to monitor and drive Claude Code from your phone. Because it ships from the vendor, it has the tightest possible integration with Claude Code. It is built for Claude Code specifically.
Remote Control is Claude-only and first-party. Pushary is the cross-vendor layer: one install, one policy and one audit trail for the agents it hooks into (Claude Code, Codex, Cursor and Gemini CLI). Hermes and other MCP clients can ask you through the same app. Remote Control uses Claude Code's own permission rules, which cover Claude Code alone; Pushary applies one set of rules across those agents and machines. It also records every question and decision (export on Agent Pro and above), sends approvals to Slack, and has one Stop all for every hook-connected agent. If your whole workflow is a single Claude Code session, Remote Control is the most native option. If you run more than Claude, or need one policy and one record across agents, that is what Pushary is for.
No. They are independent. You could use Anthropic's Remote Control for a deep single-session Claude experience and Pushary for the cross-vendor pieces: one policy across agents, Stop all, Slack, and an audit trail that spans the agents and machines you connect.
First-party wins on depth and support for its own tool. If you only run Claude Code and want the most native control surface, backed by the vendor, Remote Control is the right pick. Pushary makes the opposite bet: be vendor-neutral, and add the policy and audit layer that a single-vendor tool will not build for its competitors.
On the web, Pushary's Agent plan is $9.99/mo with a 3-day free trial. The paid line funds the policy engine, the immutable audit trail, and cross-vendor support beyond Claude Code.
Set your guardrails once for Claude Code, Codex, Cursor and Gemini CLI, approve from your phone, and keep a record of every decision. Hermes and other MCP clients can ask you through the same app. 3-day free trial on the web.
Start your trial