# GPT agent platforms with approval controls

> Source: https://theagentsindex.com/collections/gpt-agent-platforms-with-approval-controls (curated, re-verified)

Letta's Pro plan costs $20 a month, the smallest US dollar paid plan among these ten agent platforms. n8n's Cloud Starter runs from €20 a month (about $22). GC AI has the highest US dollar figure at $500 a seat a month. For ElevenLabs Agents and LuMay, see current pricing. Moveworks has no public list price and is quoted by sales.

All ten run GPT models and let you choose what an agent may do before it asks a person. They differ in how finely you can choose. GC AI splits grants by action type, so searching an inbox and sending mail are separate grants. Lindy sets a guardrail on every connected account, and reads are never gated. LangChain's Fleet agents can run as themselves or as the calling user.

Model choice splits them too. Moveworks is the one member without Claude support. The other nine support Claude as well as GPT. For running on your own infrastructure, n8n and LuMay both offer self-hosting.

| Name | What it is | Why it's here | Pricing | API |
| --- | --- | --- | --- | --- |
| [ElevenLabs Agents](https://theagentsindex.com/elevenlabs-conversational-ai.md) | Hosted platform for building voice and chat agents, with telephony, SDKs and a choice of LLM. | Voice is the reason to pick ElevenLabs Agents. Tool use is gated per MCP server and per tool, from always ask to no approval at all. The vendor says OpenAI models stay fully available under EU data residency. Telephony and CRM work are yours to assemble. For price, see current pricing. | Free, then $6 a month | Yes |
| [GC AI](https://theagentsindex.com/gc-ai.md) | Legal AI platform for in-house counsel, spanning drafting, contract review, playbooks and research. | At $500 a seat a month, GC AI is built for in-house legal teams. A card in the chat asks for approval before anything leaves. The vendor states a zero data retention agreement with OpenAI. US Case Law research costs extra on the Individual plan. | From $500 a seat a month | Yes |
| [Genspark](https://theagentsindex.com/genspark.md) | All-in-one AI workspace with an autonomous Super Agent plus agents for slides, docs, media and code. | For $24.99 a month on Plus, Genspark puts research and slides in one web workspace. Its agents run long multi-step work alone, and the vendor describes human-in-the-loop checkpoints where a person approves the steps that matter. Credit use varies by task, and unused monthly credits do not roll over. | Free, then $24.99 a month | Yes |
| [Gumloop](https://theagentsindex.com/gumloop.md) | Multiplayer AI agent builder: teams build agents over their apps while IT keeps governance and spend control. | Made for teams building agents over their own apps, Gumloop starts at $37 a month and lets IT keep governance and spend control. Each connector has an approval mode, and a middle setting leaves reads free while gating writes. Several GPT models are served through OpenAI directly. Free use is limited. | From $37 a month | Yes |
| [LangChain](https://theagentsindex.com/langchain.md) | Open source agent frameworks plus LangSmith, a hosted platform to trace, evaluate, deploy and govern agents. | LangChain and LangGraph themselves are free and open source, and LangSmith Plus costs $39 a seat a month. Human-in-the-loop interrupts gate agent actions, and a first-party langchain-openai package covers GPT and Azure OpenAI. The abstractions have churned across versions, so expect a learning curve. | Free, then $39 a seat a month | Yes |
| [Letta](https://theagentsindex.com/letta.md) | Stateful agent harness, cloud and SDK for agents with persistent, git-versioned memory. | Persistent memory is Letta's draw, and Pro costs $20 a month for up to 20 agents. Four permission modes run from strict to unrestricted, with allow and deny patterns for tools. OpenAI API keys and ChatGPT Plus or Pro Codex subscriptions both connect. Its backing is a single seed round from 2024. | Free, then $20 a month | Yes |
| [Lindy](https://theagentsindex.com/lindy.md) | Slack-native AI teammate that connects to company tools and does delegated work, billed per seat with pooled credits. | Slack is home for Lindy, which joins channels as a teammate, and Plus costs $29.99 a user a month with 3,000 credits. Approval is enforced in shared Slack threads, while DMs and web chat are not. Credit use runs from 2 to 2,500 credits a task, which makes a busy workspace's bill hard to predict. | From $49.99 a seat a month | No |
| [LuMay](https://theagentsindex.com/lumay.md) | Enterprise agentic AI platform with packaged voice, legal, compliance and workflow agents; SaaS, on-prem or air-gapped. | On-prem and air-gapped installs sit beside SaaS for LuMay, which ships packaged agents for voice and legal work. Human-in-the-loop gates sit in its orchestration layer, so an agent's reach before it asks is a policy you configure. The hosted default routes through Azure OpenAI inside Microsoft's enterprise boundary. Its price: see current pricing. | From $249 a month | Yes |
| [Moveworks](https://theagentsindex.com/moveworks.md) | Enterprise AI assistant and low-code agent platform for employee search and task automation. | Large, often regulated enterprises are Moveworks' audience, with one assistant handling IT and HR requests. Policy validators enforce rules before an action runs, and an approvals engine can require human sign-off in sequence or in parallel. GPT is the only model family selectable in LLM Actions. Sales quotes every deal. | From $150 a seat a year | Yes |
| [n8n](https://theagentsindex.com/n8n.md) | Workflow automation platform for building and running AI agents on a visual canvas, with code where needed. | Visual canvas building is n8n's pitch, and self-hosted Community is free while Cloud Starter runs from €20 a month (about $22) for 2,500 executions. Approval is wired per tool, with a reviewer answering in Slack or Gmail, and a tool without a review step runs unattended. The licence is fair-code, source-available rather than OSI open source. | From €24 a month (about $27) | Yes |

## What it takes to be on this list

Each platform here does two things a buyer can check on its own site. It runs OpenAI's GPT models. It also lets you set which actions an agent takes alone and which wait for a person's approval. The other agent platforms in this category do one of these at most. Entries run A to Z by name. Before buying, find out where approval is enforced on the plan you would buy, since some gates cover only certain channels or tools. Also confirm which plan carries the GPT models you want and how usage is metered.