Pushary
Sends AI agent permission requests to a user's phone, Slack, or web app for one-tap approval or denial, with policy-based auto-approval and audit trail.
Pushary is a human-in-the-loop approval platform for AI agents, solving the problem of agents stalling on permission requests. It sends decision requests to a user's phone, Slack, or web app, allowing one-tap approval or denial. The product targets developers, IT teams, and enterprises using AI coding agents like Claude Code, Codex, and Cursor. It is positioned as a control panel that centralizes approvals across multiple agents, with policy-based auto-approval for safe actions and escalation for risky ones, delivered as a SaaS with on-premise code execution.
Key features
- One-tap approve or deny from phone
- Works with Claude Code, Codex, Cursor
- Policy-based auto-approval for safe actions
- Escalation of risky actions to phone
- Audit trail of all decisions
- Slack and n8n integrations
- Multi-agent control panel
- Set timeout and timeout actions
- X1.5/day
- Reddit0.2/day
GTM channels
- Blog
- Partner program
- Marketplace
- Community
- API
- Docs
- Changelog
ICP
- Software developers
- Engineering teams
- Enterprises