# Rippletide Docs

> Connect existing agents, inspect runtime evidence, and apply reviewed business rules at action and response boundaries.

- [What is Rippletide?](https://docs.rippletide.com/index.md): Connect an existing agent, understand its actions, and apply rules with runtime evidence.
- [Quickstart](https://docs.rippletide.com/docs/quickstart.md): Connect an agent, verify a tool-using run and response delivery, then observe your first Rule.
- [Connect an agent](https://docs.rippletide.com/docs/connect-agent.md): Instrument an existing JavaScript or TypeScript agent through its generated setup prompt.
- [Connect a Python agent](https://docs.rippletide.com/docs/connect-python.md): Install the Python SDK, apply the generated setup prompt, and verify your first live turn.
- [Connect an MCP server](https://docs.rippletide.com/docs/connect-mcp-server.md): Declare a JavaScript or TypeScript MCP server and guard its real tool dispatcher.
- [Connect OpenAI Codex](https://docs.rippletide.com/docs/connect-codex.md): Install local Codex hooks, choose what to record, and understand the limits of tool governance.
- [Write rules](https://docs.rippletide.com/docs/write-rules.md): Choose a boundary and scope, review required setup, and create a versioned Rule.
- [Business context](https://docs.rippletide.com/docs/business-context.md): Connect policy sources, propose Rules from documents and traces, and review the impact of updates.
- [Test your rules](https://docs.rippletide.com/docs/evaluate-improve.md): Test matching and non-matching cases, inspect the applied outcome, and review replay limits.
- [Observe & enforce](https://docs.rippletide.com/docs/observe-and-enforce.md): Understand Rule modes, Allow/Block decisions, and the evidence of an applied outcome.
- [Filter tool results](https://docs.rippletide.com/docs/result-filtering.md): Apply Rules to each item in a declared collection before delivering the tool result.
- [Fleet overview](https://docs.rippletide.com/docs/fleet.md): Prioritize incomplete connections, risky action calls and repeated human interventions.
- [Context & actions](https://docs.rippletide.com/docs/harness-management.md): Understand the declared agent, inspect action risks, and choose where Rules would help.
- [Traces](https://docs.rippletide.com/docs/runtime-events.md): Read runs, tool calls, policy decisions and reported outcomes without overstating the evidence.
- [Workspaces](https://docs.rippletide.com/docs/workspaces.md): Share agents, rules, and evidence with your team.
- [Privacy & capture](https://docs.rippletide.com/docs/privacy.md): Choose what runtime content to export and distinguish traces, inventory and decision audit data.
- [CLI reference](https://docs.rippletide.com/docs/cli.md): Connect, inspect traces, manage Rules and sources, and administer authorized workspaces.
- [Policy API](https://docs.rippletide.com/docs/policy-api.md): Call scoped decision endpoints and report what your application actually applied.
- [API keys](https://docs.rippletide.com/docs/api-keys.md): Use workspace management credentials and connection-scoped runtime keys in the right place.
- [Manual TypeScript setup](https://docs.rippletide.com/docs/manual-setup.md): Declare the real inventory and protect tool execution and response delivery inside a causal run.
- [Manual Python setup](https://docs.rippletide.com/docs/manual-python-setup.md): Instrument synchronous Python model, tool, and response call sites yourself.


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