> ## Documentation Index > Fetch the complete documentation index at: https://docs.godizzy.dev/llms.txt > Use this file to discover all available pages before exploring further. # Introduction > GoDizzy is a programmable API gateway for mocking, proxying, and stress-testing agent tool calls without changing client code. GoDizzy sits in front of your agent's HTTP tool calls. Your agent talks to one stable GoDizzy URL, and GoDizzy decides whether to return a mock response or proxy the request to a real upstream. ## Why teams use GoDizzy AI agents depend on APIs that fail in messy ways: timeouts, malformed payloads, partial results, schema drift, and rate limits. GoDizzy lets you reproduce those conditions on demand without rewriting the agent. Every route collection gets its own GoDizzy URL. Your agent keeps calling that URL while you change behavior in the dashboard. Each routing rule can return a fixture or forward to a live backend. Simulate slow responses, broken schemas, and error codes before they happen in production. Most routing changes happen in GoDizzy, not in your agent config or application code. ## Core building blocks **Environments** — GoDizzy uses a shared **production** environment and a private **development** environment per userGoDizzy uses a shared **production** environment and a private **development** environment per user. **Environments** — GoDizzy uses a shared **production** environment and a private **development** environment per user. **Route collections** — A collection is a gateway with a generated subdomain, a target endpoint, and a set of routing rules. **Routing rules** — Rules match by HTTP method and path, then either mock or proxy. Higher-priority rules win first. **Mock responses** — Return deterministic JSON fixtures, custom status codes, headers, and optional latency. **Proxy response shaping** — Forward to a real upstream and optionally override status, headers, or JSON body fields before the response goes back to your agent. **Mock webhooks** — Fire outbound callbacks after serving a mock response so you can test async integrations. **MCP** — GoDizzy exposes an MCP server for environments, collections, and rules. See [MCP](/guides/mcp). ## Get started Create a development environment, a collection, and your first rule. Learn the routing model, default rule behavior, and request flow. Connect an MCP client and manage environments, collections, and rules programmatically. Build deterministic fixtures with templates, latency, and rollback. AI agents depend on APIs that fail in messy ways: timeouts, malformed payloads, partial results, schema drift, and rate limits. GoDizzy lets you reproduce those conditions on demand without rewriting the agent. Every route collection gets its own GoDizzy URL. Your agent keeps calling that URL while you change behavior in the dashboard. Each routing rule can return a fixture or forward to a live backend. Simulate slow responses, broken schemas, and error codes before they happen in production. Most routing changes happen in GoDizzy, not in your agent config or application code. ## Core building blocks **Environments** — GoDizzy uses a shared **production** environment and a private **development** environment per user. GoDizzy uses a shared **production** environment and a private **development** environment per user**Route collections** — A collection is a gateway with a generated subdomain, a target endpoint, and a set of routing rules. **Routing rules** — Rules match by HTTP method and path, then either mock or proxy. Higher-priority rules win first. **Mock responses** — Return deterministic JSON fixtures, custom status codes, headers, and optional latency. **Proxy response shaping** — Forward to a real upstream and optionally override status, headers, or JSON body fields before the response goes back to your agent. **Mock webhooks** — Fire outbound callbacks after serving a mock response so you can test async integrations. **MCP** — GoDizzy exposes an MCP server for environments, collections, and rules. See [MCP](/guides/mcp). Create a development environment, a collection, and your first rule. Learn the routing model, default rule behavior, and request flow. Connect an MCP client and manage environments, collections, and rules programmatically. Build deterministic fixtures with templates, latency, and rollback.