# AgentSIM

- [Quickstart](https://docs.agentsim.dev/quickstart.md): Connect AgentSIM to an owned sign-in flow with self-serve live SMS.
- [Availability](https://docs.agentsim.dev/availability.md): Self-serve live sessions, supported channels, and current limits.
- [Concepts](https://docs.agentsim.dev/concepts.md): Challenge, policy, and verdict — and the shipped SDK and MCP names that map to them.
- [Why not Twilio?](https://docs.agentsim.dev/why-not-twilio.md): When to use AgentSIM's auth control plane instead of a raw cloud phone API.
- [Authentication](https://docs.agentsim.dev/authentication.md): How to generate, use, and rotate AgentSIM API keys.
- [Browser Access](https://docs.agentsim.dev/browser-access.md): Connect an approved browser to an exact owned origin and approve every run once.
- [SMS numbers](https://docs.agentsim.dev/number-classes.md): Number availability, delivery limits, and choosing the right kind of test.
- [Policies & Verdicts](https://docs.agentsim.dev/policies-verdicts.md): Control which services agents can authenticate with and audit all access decisions
- [Compatibility](https://docs.agentsim.dev/supported-services.md): How to verify SMS delivery for an authentication flow you own.
- [Why no SMS arrives](https://docs.agentsim.dev/troubleshooting/no-sms-arrives.md): Distinguish number rejection, delivery problems, and code extraction failures.
- [Error reference](https://docs.agentsim.dev/errors.md): Interpret API errors and decide when to retry.
- [Integrations Overview](https://docs.agentsim.dev/integrations/overview.md): Choose the right AgentSIM path: MCP, SDK, browser automation, or controlled auth on apps you own.
- [Agent runtimes](https://docs.agentsim.dev/integrations/agent-runtimes.md): Connect assistants through MCP or call the SDK from your agent runtime.
- [Browser Automation](https://docs.agentsim.dev/integrations/browser-automation.md): Use AgentSIM with Playwright, Cypress, Browserbase, Stagehand, Selenium, and browser-agent QA flows.
- [Controlled Auth Workflows](https://docs.agentsim.dev/integrations/controlled-auth.md): Use AgentSIM with Supabase Auth, Auth0 Passwordless SMS, Firebase Phone Auth, Clerk, and custom OTP systems you control.
- [Agent Skill](https://docs.agentsim.dev/integrations/agent-skill.md): Install or reference the AgentSIM OTP Sessions Agent Skill so agents use AgentSIM safely and diagnose delivery outcomes.
- [API Overview](https://docs.agentsim.dev/api-reference/overview.md): Base URL, authentication, versioning, and conventions for the AgentSIM REST API.
- [Identify Agent](https://docs.agentsim.dev/api-reference/identify-agent.md): POST /v1/agents/identify — Add an Agent record without opening a challenge.
- [Open Challenge](https://docs.agentsim.dev/api-reference/provision.md): Open a live SMS session or an explicit test-only channel.
- [Inject Test Email](https://docs.agentsim.dev/api-reference/inject-email.md): Parse an injected email body in a test session. Inboxes cannot receive email.
- [Wait for Verdict](https://docs.agentsim.dev/api-reference/wait-for-otp.md): POST /v1/sessions/:id/wait to wait for a challenge verdict.
- [Get Messages](https://docs.agentsim.dev/api-reference/get-messages.md): Read SMS message records and extracted codes for a session.
- [Close Challenge](https://docs.agentsim.dev/api-reference/release.md): DELETE /v1/sessions/:id — Close the challenge session.
- [List Challenges](https://docs.agentsim.dev/api-reference/list-numbers.md): GET /v1/sessions — List challenge sessions for your account.
- [Webhooks](https://docs.agentsim.dev/api-reference/webhooks.md): Receive parsed SMS codes at your application endpoint.
- [Python SDK](https://docs.agentsim.dev/sdk/python.md): The official AgentSIM Python SDK: open_challenge opens an SMS challenge, wait_for_verdict waits for the verdict.
- [TypeScript SDK](https://docs.agentsim.dev/sdk/typescript.md): The official AgentSIM TypeScript SDK: openChallenge opens an SMS challenge, waitForVerdict waits for the verdict.
- [MCP Overview](https://docs.agentsim.dev/mcp/overview.md): Use AgentSIM from Claude, Cursor, Windsurf, and any MCP-compatible assistant — open_challenge, wait_for_verdict.
- [Connect](https://docs.agentsim.dev/mcp/connect.md): Set up AgentSIM's MCP server in Codex CLI, Claude Desktop, Claude Code, Cursor, or another MCP client.
- [Tools Reference](https://docs.agentsim.dev/mcp/tools.md): MCP tools: identify_agent, open_challenge, wait_for_verdict, and legacy aliases.
- [Workflow Recipes](https://docs.agentsim.dev/mcp/recipes.md): Prompts for AgentSIM MCP: open a challenge, wait for a verdict, close the session.
- [Troubleshooting](https://docs.agentsim.dev/mcp/troubleshooting.md): Common MCP connection issues and their fixes.
- [Changelog](https://docs.agentsim.dev/changelog.md): What's new in AgentSIM — releases, features, and fixes.

## OpenAPI Specs

- [openapi](/openapi.json)

## Optional

- [Console](https://console.agentsim.dev)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.