One account that gives AI agents email, phone, payments, post and an off switch
- Remote
- Free trial
- Agent approvals
One account that gives AI agents email, phone, payments, post and an off switch
Lets AI agents put questions and tasks on your list, answered with one tap
Not checked yetLets AI agents use your API keys through a local proxy without ever seeing them
Runs locallyOpen-source workspace where AI agents run product work and humans set direction
Safety gateway that checks MCP tool calls and holds risky ones for approval
Not checked yet
Pre-trade risk checks for AI agent orders against your own limits
Not checked yet
Build, test, deploy and operate AI agents hosted in the EU
Not checked yet
Auto-approve small agent actions under your rules; ask you about the rest
Not checked yetMCP server that screens a software or AI project with 11 questions before investing
Not checked yetMCP server exposing a command approval gate and classify/route/score tools for coding
Runs locallyPolicy checks and tool-trust verification for AI agent and MCP teams before tools run
Not checked yetLocal pre-action safety check for coding agents that scores risky commands before they run
Runs locallyLets AI agents ask a named person for approval, a decision or an answer, with an audit
Not checked yetHuman approvals, autonomy limits and an audit log for service-firm agents
Not checked yetPermissioned, audited AI access to company computers and projects
Not checked yetHuman approval gate for AI agents, decided on the approver's phone with a signed record
Runs locallyStatic scanner that audits other MCP servers' tool code for risky patterns
Runs locallyOpen-source gate that lets AI agents read freely but blocks unapproved writes to MCP
Runs locallyNegotiation rooms where B2B agents bargain under owner rules and lock deals
Not checked yetAgent-to-agent B2B negotiation with owner rules and locked deals
Not checked yetHolds AI-agent actions until live facts agree or a person decides
Not checked yetAI managers that audit and optimise ad accounts, with approval-gated changes
Not checked yetPer-client keys, budgets and approvals for agents, run on your own machine
Not checked yetRead-only CLI and MCP server that audits agent tool setups and drafts least-privilege
Runs locallyPolicy gate that authorizes AI agent actions before they execute
Not checked yetLocal MCP proxy that checks each tool call against Agent Trust policy before it runs
Runs locallyPolicy, human approval and read-back verification with a tamper-evident ledger for AI
Runs locallyPre-execution policy gate that returns allow, review or block decisions for AI agent
Not checked yetMCP operator server for answering held CTRLRun approvals from an AI assistant
Not checked yetPaid x402 evaluator that checks whether evidence and authority warrant an agent's action
Not checked yetSee every kind of AI for agent approvals: tools, agents, MCP servers and skills.Ranked: the best AI for agent approvals.
The most popular MCP servers for agent approvals on AI-Powered right now are STACK, AskHooman, vernam secrets broker, Maskin and Halto. Rankings combine upvotes, recent activity and, for open-source projects, GitHub stars.
Yes. Of the 83 MCP servers for agent approvals listed, 35 are free or open source and 24 have a free plan. Use the Pricing filter to see only those.
AI-Powered lists 83 MCP servers for agent approvals. New listings are added daily from the official MCP Registry, GitHub and maker submissions.
Open the server's page and copy the snippet for your client. Claude Code uses a single "claude mcp add" command; Claude Desktop, Cursor and VS Code use a JSON entry in their MCP config file. Remote servers only need their URL.