Home
API Playground
SearchContentsAnswerAgentBETAWebsetsMonitors
Management
UsageBillingAPI KeysTeam SettingsWebsets Settings
Learn
DocsExa MCPTemplates

Generated Page

Programmatic SEO pages grounded in fresh source material

Live preview

Checking API key

How it works

Each page is one request to Exa's /search endpoint with type: "deep". The query asks Exa to research the keyword, then the systemPrompt and outputSchema shape the result into a page.

The variants differ by schema. A listicle returns items and faqs. A comparison returns exactly two contenders and a verdict. Alternatives returns ranked replacements. FAQ returns only question and answer pairs.

Sources come from output.grounding. The preview keeps inline markdown links in prose and renders a compact favicon source row deduped by domain.

Source code

npm install exa-js
import Exa from "exa-js";
const exa = new Exa("YOUR_API_KEY");

const keyword = "Top 10 AI sales tools for B2B SaaS";

const result = await exa.search(
  `Research the best options for: ${keyword}. Find recent rankings, reviews, and comparisons.`,
  {
    type: "deep",
    numResults: 12,
    systemPrompt: `Generate a programmatic SEO listicle targeting "${keyword}". Rank real products by fit for the named audience. Plain prose, grounded claims, natural inline links, no hype.`,
    outputSchema: {
      type: "object",
      required: ["metaTitle", "metaDescription", "intro", "items", "faqs"],
      properties: {
        metaTitle: { type: "string" },
        metaDescription: { type: "string" },
        intro: { type: "string" },
        items: {
          type: "array",
          items: {
            type: "object",
            required: ["name", "url", "summary"],
            properties: {
              name: { type: "string" },
              url: { type: "string" },
              summary: { type: "string" },
            },
          },
        },
        faqs: {
          type: "array",
          items: {
            type: "object",
            properties: {
              question: { type: "string" },
              answer: { type: "string" },
            },
          },
        },
      },
    },
    contents: { highlights: true },
  }
);

const page = result.output.content;
const citations = result.output.grounding.flatMap((g) => g.citations);
Ask ExaBot