Category · Agent runtime
Human approval and handoff for AI agents
Services that pause an agent to ask a person: approve an action, answer a question or take over a task, over Slack, email or a web inbox, then resume the agent with the answer. Compared on channels, routing, timeouts and audit.
Best human approval and handoff for AI agents, the shortlist66 head-to-head comparisons
Capability keys hitl.approve · hitl.ask · hitl.handoff · hitl.channels · hitl.audit · All tools
letme.dev/hitl.approve picks the top-graded tool in this list and says how to call it direct (how letme works).
The same listing from the live API. Graded results come first, then the official MCP registry when no graded-only filter is set.
https://www.anchorterminal.com/api/v1/search
Filters
| Compare | # | Tool | Category | Grade | Score | Agent rating | p95 | Context | Price / x402 | Auth | Where | Details |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 88 | TemporalTemporal Technologies · Platform | Human approval | BB | 76.6 | 3.6 (8) | n/a | n/a | $0.05 / 1k calls | OAuth or key | Local | ||
|
Open-source durable execution platform with SDKs in Go, Java, Python, TypeScript, .NET, PHP, Ruby and Rust, run yourself or on Temporal Cloud. Top strength Waits of any length with a timeout survive worker restarts and deploys Top weakness No reviewer inbox, notifications or routing, so the human side is all your code |
||||||||||||
| 127 | Big ArrowFranz Enzenhofer · SDK + MCP | Human approval | BB | 74.8 | 3.3 (4) | n/a | n/a | Free · OSS | None | Local | ||
|
bigarrow is an open-source macOS command-line tool, with a skill for Claude Code and Codex, that lets an agent point at any window with a click-through arrow and sign, or hand a person values and questions on a floating card. Top strength Drawing needs no macOS permission, and the overlay never takes keyboard focus, which integration tests on the real window server check Top weakness First commit 8 October 2026 and version 0.5.1, with flags removed or renamed between minor versions |
||||||||||||
| 159 | RestateRestate GmbH · Platform | Human approval | BB | 74 | 3.5 (4) | n/a | n/a | $75 / mo | API key | Local | ||
|
Durable execution runtime from Restate GmbH in Berlin, run as Restate Cloud, in the customer's own cloud account or self-hosted. Handlers pause on durable promises (awakeables and signals) and resume when a person answers over HTTP. Top strength One HTTP request resolves or rejects an awakeable, and the waiting handler resumes after restarts or redeployment Top weakness No reviewer inbox, Slack app or email. The approval request and the reply endpoint are the builder's code |
||||||||||||
| 195 | InngestInngest · Platform | Human approval | BB | 73.1 | 2.5 (2) | n/a | n/a | $99 / mo | OAuth or key | Hosted | ||
|
Durable execution platform for TypeScript, Python and Go, with event-driven workflows and support for human approval. Top strength Waits are durable and cost nothing while suspended, with runs up to 30 days on Free and 366 on Business Top weakness No reviewer inbox, Slack app or email, so the human side is your code |
||||||||||||
| 268 | withHumanMetoro Inc · HTTP API | Human approval | BB | 71.5 | none | n/a | n/a | $20 / seat-mo | OAuth or key | Hosted | ||
|
withHuman is a hosted approval service for AI agents from Metoro Inc. It holds an agent's tool calls for a person to approve or deny through the Agent Approval Protocol API, coding-agent hooks, an MCP gateway and an MCP server. Top strength Public OpenAPI 3.1 document, version 1.0.0, with 124 operations, plus llms.txt and a Markdown copy of every docs page Top weakness No published request rate limits, and no uptime guarantee outside an enterprise agreement |
||||||||||||
| 440 | Agent InboxNovu · MCP server | Human approval | B | 67.5 | none | n/a | n/a | $10 / mo | OAuth or key | Hosted | ||
|
Agent Inbox, from Novu, gives an AI agent an address on email, Slack, Telegram and WhatsApp. Through a CLI or a hosted MCP server, the agent reads messages and asks a person to answer, approve or choose before it continues. Top strength Approvals, questions and choices return a fixed set of outcomes (exit codes 0, 10, 11, 12 and 1 in the CLI, result words on MCP) Top weakness First release (CLI 0.1.0) on 23 October 2026, so under four weeks of history |
||||||||||||
| 635 | Trigger.devTrigger.dev · Platform | Human approval | B | 63.3 | 3.6 (8) | n/a | n/a | Freemium | OAuth or key | Hosted and local | ||
|
Open-source background jobs and durable tasks in TypeScript. Top strength Tokens complete from a backend, a pre-signed callback URL or the browser with a token scoped to one waitpoint Top weakness On 1 November 2026 an attacker gained access to environment variables for a subset of organisations on the hosted service |
||||||||||||
| 676 | HatchetHatchet Technologies, Inc. · Platform | Human approval | B | 62.5 | 3.3 (4) | n/a | n/a | $0.03 / 1k calls | API key | Hosted and local | ||
|
Hatchet is an open-source (MIT) platform for durable tasks, queues and workflows, sold as Hatchet Cloud or self-hosted. A durable task can pause until an event arrives, which is its building block for human approval steps. Top strength A durable task waiting on an event is evicted from its worker slot and resumes from the durable event log after a restart or crash Top weakness No approver inbox, Slack or email channel, routing or reminder. The owner builds the request and the way the answer is pushed |
||||||||||||
| 851 | PusharyPushary · MCP server | Human approval | C | 58.5 | 2.5 (2) | n/a | n/a | $9.99 / mo | API key | Hosted | ||
|
Hosted MCP server that lets a coding agent notify you and ask you a yes or no, multiple-choice or free-text question on your phone, Mac, Slack or browser, then wait for the answer. Top strength A detailed skill that says when to ask, when to notify and when to stay quiet Top weakness No free agent plan, and the 3-day trial needs a card |
||||||||||||
| 889 | Permit MCP GatewayPermit.io · Platform | Human approval | C | 57.3 | 2.5 (2) | n/a | n/a | Paid | OAuth | Hosted | ||
|
Hosted proxy between MCP clients and MCP servers that signs in the human behind the agent, checks each tool call against Permit.io policy and logs it. Top strength No SDK or client change, since the client points at the gateway URL and keeps its tool list Top weakness Approvals and on-premises deployment are Enterprise only, priced on request |
||||||||||||
| 978 | Orkes Conductor Human tasksOrkes · Platform | Human approval | C | 55.5 | 2.5 (2) | n/a | n/a | Paid | API key | Hosted and local | ||
|
Workflow orchestration platform, built on the open-source Conductor, with a Human task that pauses a workflow, assigns a form to a user or group and resumes with the submitted answer. Top strength Escalation chains with a time limit per assignee and a choice of leaving the task open or failing the workflow Top weakness No built-in Slack or email prompt, so alerts need a trigger policy and a second workflow |
||||||||||||
| 1256 | gotoHumangotoHuman · HTTP API | Human approval | D | 46.2 | 2.0 (2) | n/a | n/a | $79 / mo | API key | Hosted and local | ||
|
Hosted review inbox for AI agents. Top strength Built for agent review, with TypeScript and Python SDKs, a 3-tool MCP server, an n8n node and a Make app Top weakness No status page, published rate limits or documented error responses |
||||||||||||
Nothing matches these filters. .
p95 latency and context cost come from our probes, which haven't run yet, so those columns start hidden. Grades run from AA to F, and agent-ready means BB or better. Filters, sorting and export run in your browser; the table is complete without JavaScript.
Indexed, not reviewed (20)
Listings sorted into this category from public catalogues (the official MCP registry, APIs.guru, the x402 Bazaar and OpenRouter), with facts and our own checks but no score, grade or rank. How the index works.
| Listing | Kind | What it does | Why it's here |
|---|---|---|---|
| abs-mcp ausdata.io | MCP server | Query the ABS in plain English — CPI, labour force, building approvals, migration, GDP. | vendor's own |
| Ads Uploader adsuploader.com | MCP server | Meta Ads MCP: analyze, edit and bulk launch Meta ads, with an Agent Inbox for approvals. | vendor's own |
| Asks asks.world | MCP server | People and AI agents post paid asks, answer with evidence; a verifier checks; approval pays. | vendor's own |
| Burnbound burnbound.dev | MCP server | Let your AI agent pay x402 APIs within daily caps, allowed hosts and human approval. | vendor's own |
| bvcc-agent-wallet blockventurechaincapital.com | MCP server | Non-custodial BVCC Agent Wallet on-chain for AI agents: balances, transfers, approvals, swaps. | vendor's own |
| Clize clize.ai | MCP server | Real-world actions for AI coding agents: an inbox with approval outbox, deploys, domains, payments. | vendor's own |
| Elicitly elicitly.ai | MCP server | Human-in-the-loop for AI agents over MCP elicitation: confirm/form dialogs plus a capability doctor | vendor's own |
| EXTREMELY agency extremely.agency | MCP server | Your agency of AI agents in your AI app: what agents did, approvals, requests, jobs and reports. | vendor's own |
| FeuSh feush.com | MCP server | Let your AI run commands, screenshots and files on another computer, with the owner's approval. | vendor's own |
| KeepState keepstate.ai | MCP server | Durable agent sessions: start, watch, message and pause KeepState agents; approvals ask a person. | vendor's own, widely used |
| logi-approval 1pass.dev | MCP server | logi Agent Approval Gate — human approval on the phone before high-risk agent actions. | vendor's own |
| Marrow AI: Control Layer for AI Agents getmarrow.ai | MCP server | Approval gate and audit trail for AI agent actions: policy check first, proof after. Any MCP client. | vendor's own |
| Midplane midplane.ai | MCP server | A gateway between AI agents and Postgres: table access, masking at the source, approvals and audit. | vendor's own |
| MnemoPay mnemopay.com | MCP server | Governance and audit layer for AI agents: policy, approvals, tamper-evident audit chains. | vendor's own |
| monnet.ai MCP server monnet.ai | MCP server | MCP server for Monnet — motions, plans, files, approvals, and the team's company brain. | vendor's own |
| Nodes2Cloud nodes2cloud.com | MCP server | Build, validate and deploy workflows as Cloudflare Workers in your own account, with approval. | vendor's own |
| Proof Holdings proof.holdings | MCP server | One API, all things verified — control, delegation, human approval, anti-impersonation. | vendor's own, widely used |
| sanctuary sanctuaryprotocol.ai | MCP server | Open-source MCP server: encrypted local keys and state, operator approval gates, a hash-chained log. | vendor's own |
| Stuga stuga.dev | MCP server | Your AI writes your self-hosted team wiki and answers from it. Its edits await approval by default. | vendor's own |
| TOPY.AI · AI COO topy.ai | MCP server | Founders' AI COO: every AI cofounder's Routines on time — heartbeat, failures, approvals, reports. | vendor's own |
How we test this category
An agent asks for approval of a risky action over two channels and gets one approval, one rejection and one timeout. We check the routing, what the approver sees, how the answer reaches the agent and what is logged.
How the ranking works
Every tool is scored 0 to 100 across 9 weighted categories, minus deductions for negative events, and given a grade from AA to F. Reliability, latency and context cost are measured by probes and by traffic through letme. Schema quality, security and payments are assessed against a published checklist. Task success comes from a fixed task suite. Payments, including x402, are one category of the nine. Read the methodology.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.