# AgentSIM

> Give your agents a safe path through the auth challenges you own.

## What It Does

AgentSIM is a control plane for agents working through authentication challenges on verified services they control. It verifies the service origin, applies policy, and returns a structured SMS verdict. Email and magic-link paths are test-only. Passkeys and CAPTCHA stop safely. Google and Stripe are refused.

When a browser agent or CI job hits an auth wall, AgentSIM runs the challenge, applies your policy, and returns a verdict so the run can continue. Agents can use it through SDKs, REST, webhooks, or MCP.

- **Owned auth, CI/QA, and agent runs.** Challenge coordination for apps you own, test harnesses, and live agent runs — not a phone shop, not a bypass.
- **Policy verdicts.** When a target is refused (Google, Stripe by denylist) or nothing arrives, the session records the outcome so the run can branch.

One live session covers a successful number assignment or extension for one verified service origin. Hobby includes 10 shared live sessions per UTC month. Builder is $99/month and includes 50 shared live sessions. Live SMS is available now for verified US service origins; Browser Access uses the same allowance when enabled. Review tax, renewal, and any existing agreement at checkout before payment.

## Authorized Use Only

AgentSIM is for **authorized testing of systems you own or are permitted to test** and for auth workflows on providers you control.

**Not for** creating accounts on third-party services you do not own, bypassing anti-abuse systems at Google, Meta, Stripe, or banks, bulk account creation, ban evasion, or any use that violates a target service's terms or applicable law. Google and Stripe are refused by policy denylist. Not a phone/OTP rental shop.

## Compatibility

Use an app you own with a real public HTTPS origin. Live SMS requires automated verification of the exact service origin in Agent control, a matching allow policy, remaining allowance, and a ready shared US line. Google and Stripe are refused by policy. Test-only injected messages do not prove carrier delivery.

Current access and channel support: https://docs.agentsim.dev/availability
Compatibility checks: https://docs.agentsim.dev/supported-services


## Quick Start

Primary published SDK (TypeScript). `openChallenge` opens an SMS challenge session. `waitForVerdict` waits for the verdict. `provision` / `waitForOtp` are aliases that still work. Timeout is seconds, not milliseconds.

```bash
bun add @agentsim/sdk
# or
pip install agentsim-sdk
```

```typescript
import { openChallenge } from "@agentsim/sdk";

// Replace this target with your owned public HTTPS origin. Requires live access.
await using num = await openChallenge({ agentId: "checkout-bot", serviceUrl: "https://staging.example.com" });
// Enter num.number in your app and request its SMS code.
const otp = await num.waitForVerdict({ timeout: 60 });
// otp.otpCode is the parsed SMS verdict on apps you own
```

Python ships the same REST operations as `agentsim.open_challenge` / `wait_for_verdict` (no `AgentSIM()` class). `provision` / `wait_for_otp` are aliases. MCP uses `identify_agent` to add a console agent without a challenge, then `open_challenge` / `wait_for_verdict`; `provision_number` / `wait_for_otp` are aliases.

## Product

- [CI / QA](https://agentsim.dev/use-cases/browser-agent-otp-qa): CI/QA for browser agents hitting auth walls on apps you own
- [How It Works](https://agentsim.dev/#how-it-works): Challenge · Policy · Verdict
- [Use Cases](https://agentsim.dev/#use-cases): Owned auth, CI/QA, agent runs
- [Agent Skill](https://agentsim.dev/.well-known/agent-skills/otp-sessions/SKILL.md): AgentSIM challenge sessions skill for compatible agent clients
- [Agent Landing Page](https://agentsim.dev/agents): Agent-facing overview of auth.md registration, scoped bearer tokens, JWT refresh, and revocation
- [Pricing](https://agentsim.dev/#pricing): Hobby includes 10 shared live sessions per UTC month. Builder is $99/month and includes 50 shared live sessions. Live SMS is available now for verified US service origins; Browser Access uses the same allowance when enabled. Exact terms at https://console.agentsim.dev/billing
- [Sign Up](https://agentsim.dev/sign-up): Create an account and start with the free live SMS allowance

## Documentation

- [Quickstart](https://docs.agentsim.dev/quickstart): First challenge session
- [Concepts](https://docs.agentsim.dev/concepts): Challenge, policy, and verdict model
- [Authentication](https://docs.agentsim.dev/authentication): API key setup and protected API calls
- [Agent Registration](https://agentsim.dev/auth.md): auth.md service_auth registration, user claim approval, short-lived agt_ bearer tokens, and JWT-bearer refresh for agents
- [API Reference](https://docs.agentsim.dev/api-reference/overview): Identify agents, open challenges, wait for verdicts, messages, release, and webhooks
- [MCP Tools](https://docs.agentsim.dev/mcp/tools): Available MCP primitives
- [TypeScript SDK](https://docs.agentsim.dev/sdk/typescript): bun add @agentsim/sdk
- [Python SDK](https://docs.agentsim.dev/sdk/python): pip install agentsim-sdk
- [MCP Server](https://docs.agentsim.dev/mcp/connect): Claude Code, Cursor, Windsurf — one command setup

## MCP Integration

AgentSIM works where agents run: Claude Code, Cursor, Windsurf, LangChain, LlamaIndex, Vercel AI SDK, Playwright, and custom browser-agent harnesses. Use MCP, SDKs, REST, or webhooks.

Add to any MCP-compatible client:

```bash
claude mcp add agentsim -e AGENTSIM_API_KEY=YOUR_API_KEY -- npx -y @agentsim/mcp-server
```

The MCP server exposes tools: start_browser_runtime, claim_browser_runtime, create_browser_connection, open_browser_access, resume_browser_access, get_browser_session, cancel_browser_access, identify_agent, open_challenge, wait_for_verdict, provision_number, wait_for_otp, get_messages, list_numbers, release_number. Start with identify_agent, then use open_challenge and wait_for_verdict. Aliases: provision_number, wait_for_otp.

## GitHub

- [Source Code](https://github.com/agentsimdev/agentsim): Monorepo — API, SDKs, MCP server, dashboard, docs
- [Issues](https://github.com/agentsimdev/agentsim/issues)
- [MCP Registry](https://github.com/agentsimdev/agentsim/tree/main/mcp-registry)

## Performance

- Agent control: agent connection, exact-origin verification, and policy setup
- Challenge-to-verdict: waits until the wall resolves or the timeout fires
- Number country (SMS): US shared live line

## Social

- [X / Twitter](https://x.com/agentsimdev)
- [LinkedIn](https://www.linkedin.com/company/agentsim/)
- [Dev.to](https://dev.to/fathin_dosunmu)

## Legal

- [Privacy Policy](https://agentsim.dev/privacy)
- [Terms of Service](https://agentsim.dev/terms)
- [Cookie Policy](https://agentsim.dev/cookies)