Pushary

Approve your AI agents from your phone

Pushary is a human-in-the-loop decision layer for AI agents. When an agent such as Claude Code, Codex, Cursor, or any MCP-compatible agent needs permission for a step, Pushary sends the question to your phone, Mac notch app, Slack, or web app, where you approve, deny, or answer. It includes per-tool permission policies with auto-approve rules for safe commands and escalation for risky ones, timeout fallback rules, an immutable cross-agent audit trail, and a kill switch. Setup is via `npx pushary@latest setup`. A Partner plan lets product teams embed approvals in their own agents via a decisions API with TypeScript and Python SDKs, framework adapters (Vercel AI SDK, LangGraph, LangChain, CrewAI, Mastra, OpenAI Agents SDK, Pydantic AI, and others), and signed webhooks.

Category: AI agents & frameworks
Pricing: freemium · from $9.99/mo
Free tier: no
API: yes
MCP server: yes
SDK: yes
Open source: no

#human-in-the-loop #ai-agents #approvals #mcp #notifications #audit-trail #permission-policies

Alternatives to Pushary →

Listed 2026-10-25 · last verified 2026-10-25 · extracted by @cf/zai-org/glm-5.3-flash · confidence 0.95

Your product? Claim this listing · Report a problem · Badge

API

curl https://toolgrep.com/api/products/pushary