// This script sends the document title and location to the parent window
// whenever the title changes or the URL changes.
// This is used by the main App for the inline documentation viewer.
;(function () {
  function notifyChange() {
    window.parent.postMessage(
      {
        type: 'docs.update',
        value: {
          title: document.title,
          route:
            window.location.pathname +
            window.location.search +
            window.location.hash,
        },
      },
      '*',
    )
  }

  // Watch <title> changes
  const titleEl = document.querySelector('title')
  if (titleEl) {
    const observer = new MutationObserver(notifyChange)
    observer.observe(titleEl, { childList: true })
  }

  // Patch history.pushState
  const originalPushState = history.pushState
  history.pushState = function (...args) {
    originalPushState.apply(this, args)
    notifyChange()
  }

  // Patch history.replaceState
  const originalReplaceState = history.replaceState
  history.replaceState = function (...args) {
    originalReplaceState.apply(this, args)
    notifyChange()
  }

  // Listen to back/forward navigation
  window.addEventListener('popstate', notifyChange)

  // Initial send
  notifyChange()
})()


System messages must be at the beginning of the conversation

Learn about the rules you need to follow when using the Google provider in Latitude.

Google provider rules

Latitude Docs

Welcome to Latitude - The AI Prompt Engineering Platform

Introduction

Key terminology and concepts you need to know when using Latitude

Core Concepts

Collaborate effectively by inviting your team members to your Latitude workspace.

Invite your team

Learn the basics of writing and managing prompts in the Latitude editor.

Editor

Learn how to configure model, provider, and generation parameters for your prompts.

Configuration

Learn how to test and refine your prompts interactively in the Playground.

Playground

Manage prompt versions, track changes, and collaborate effectively with your team.

Version Control

Enables automatic execution of prompts based on specific events or schedules

Triggers

Define JSON schemas to ensure structured and validated responses from AI models.

Structured Output

Enable AI models to interact with external functions and data sources using tools.

Tool use

Tips and examples for writing effective prompts in Latitude.

Best Practices

Create autonomous agents that can use tools, reason, and complete complex tasks over multiple steps.

Agents

Learn how to use Latitude's built-in tools in your prompts.

Tools

Connect Latitude to other tools and services using Model Context Protocol (MCP) servers.

MCPs

Understand the different ways to evaluate prompt performance in Latitude.

Overview

Execute evaluations on datasets or continuously on live production logs.

Running Evaluations

Use automatically generated suggestions based on evaluation results to improve your prompts.

Prompt Suggestions

Best practices for incorporating prompt evaluation into your team's development lifecycle.

Evaluation Workflows

Learn how to create, manage, and utilize datasets for batch testing and evaluation.

Use curated datasets to prevent regressions and ensure consistent prompt performance.

Golden Datasets

Learn about the key concepts of Latitude experiments

Deploy your version-controlled prompts as API endpoints via the AI Gateway.

Publishing Prompts

Set up a local Latitude instance using Docker for development and testing.

Development Setup

Deploying and managing a self-hosted Latitude instance for production use.

Production Deployment

Get help, share feedback, and connect with the Latitude community.

Community and Support

Learn about the general structure of a PromptL prompt

Structure

Learn how to configure your PromptL prompts

Messages

Learn how to define, interpret, and interpolate variables in PromptL

Variables

Learn how to add content based on conditions in your prompts

Conditional Statements

Learn how to add multiple messages based on loop conditions in your prompts

Loops and Iteration

Learn how to install and use PromptL in your project

Quick Start

Chains and Steps are used to create multi-step prompts that can interact with the AI model in stages.

Chains and Steps

Learn how to reference other prompts in PromptL

Prompt References

Explore practical examples, advanced LLM techniques, and real-world use cases to build powerful AI applications with Latitude

Learn how to implement few-shot learning with examples to improve AI performance on specific tasks

Few-shot Prompting

Enhance AI performance by assigning specific roles, personas, and expertise areas

Role Prompting

Implement step-by-step reasoning to improve AI performance on complex problems

Chain-of-Thought (CoT)

Implement multiple branching reasoning paths to solve complex problems with Latitude

Tree of Thoughts

Learn how to use contextual prompting to create more efficient and accurate AI interactions

Contextual Prompting

Learn how to implement self-consistency to improve AI reasoning reliability through multiple sampling and majority voting

Self-Consistency

Learn how to improve AI responses by first considering general principles before tackling specific tasks

Step-Back Prompting

Learn how to combine reasoning and acting in a thought-action loop to solve complex tasks using external tools

ReAct Prompting

ReAct (Reasoning and Acting) Prompting

Learn how to implement validation guardrails to ensure AI outputs meet quality standards and safety requirements

Prompt with Guardrails

Expand your prompting skills with these advanced and specialized techniques

Other Techniques

Other Prompting Techniques

Learn how to run a prompt with the Latitude SDK

Run prompt

Learn how to run a prompt with tools using the Latitude SDK

Run a prompt with tools

Learn how to get a prompt with the Latitude SDK

Get a prompt

Learn how to create a prompt with Latitude SDK

Create a prompt

Learn how to get all prompts with Latitude SDK

Get all prompts

Learn how to create a log with the Latitude SDK

Create a log

Learn how to annotate log data with the Latitude SDK to perform HITL evaluations

Annotate log (HITL)

Learn how to pause the execution of a tool and process the data

Pause a Tool Execution

Learn how to use RAG retrieval with the Latitude SDK

RAG retrieval

Learn how to render the steps of a prompt with the Latitude SDK

Render a prompt with steps

Build a chatbot that answers weather questions using Latitude and a custom weather tool. Fun, fast, and just a bit magical.

Weather Chatbot: Ask the Clouds!

Learn how to build a prompt that thinks like a human and generates jokes.

Joke Generator

Learn how to build a Deep Search autonomous agent

Deep Search

Learn how to build an intelligent customer support agent that generates personalized email responses

Customer Support Email Generator

Build a multi-agent system for live financial insights combining web search, technical indicators, and actionable recommendations.

Stock Market Analysis Agent

Learn how to build a content moderation system that can analyze user-generated content and provide feedback on its appropriateness.

Content moderation system

Implement a comprehensive QA system for customer support responses using Rating-based LLM evaluation, Exact Match rules, and Manual review

Customer Support Quality Assurance

Specification for AI-Powered Pre-Seed Startup Analysis Tool using Latitude

Pre-Seed Startup analysis

Changelog

Get started with Latitude without writing code

Product Managers

Get started with Latitude using the TypeScript SDK

AI Engineers

Learn how to add providers into Latitude to use it in your prompts.

Providers

Common questions about the OpenAI provider

OpenAI

Common questions about Anthropic provider

Anthropic

Azure

Google

Use code-based metrics and rules to objectively evaluate prompt outputs.

Programmatic Rules

Incorporate manual reviews and direct human feedback into your evaluation workflow.

Humans-in-the-Loop

Integrate Latitude into your Node.js applications using the TypeScript SDK.

TypeScript

Integrate Latitude into your Python applications using the Python SDK.

Python

Overview of Latitude's HTTP API for programmatic access and integration.

HTTP API Reference

Learn how to access and use Latitude's API to run your prompts.

API Access

Discover how to efficiently handle and process streaming events using Latitude's API.

Streaming Events

Learn how to receive real-time notifications about events that occur in your Latitude workspace.

Webhooks

Learn how to use PromptL with different LLM providers

Getting started

Prompts

Agents

Evaluations

Datasets

Experiments

Deployment

Self-Hosting

Support

Google provider rules

System messages must be at the beginning of the conversation

Getting started

Prompts

Agents

Evaluations

Datasets

Experiments

Deployment

Self-Hosting

Support

​System messages must be at the beginning of the conversation

System messages must be at the beginning of the conversation