Open source alternative to Semrush and Ahrefs
TypeScript backlink-analysisgoogle-search-console-mcpkeyword-researchmcpseoseo-agentseo-auditseo-automationseo-skillsseo-toolssite-audit
import { z } from "zod";
import { buildProjectMeta } from "@/server/mcp/context";
import { mcpResponse } from "@/server/mcp/formatters";
import { optionalMetaOutputSchema } from "@/server/mcp/output-schemas";
import { withMcpProjectAuth } from "@/server/mcp/project-auth";
import { projectIdSchema } from "@/server/mcp/schemas";
import {
promptExplorerInputSchema,
promptExplorerModelResultSchema,
promptExplorerResultSchema,
} from "@/types/schemas/ai-search";
const inputSchema = promptExplorerInputSchema.extend({
projectId: projectIdSchema,
prompt: promptExplorerInputSchema.shape.prompt.describe(
"The exact question to ask, up to 500 characters. Preserve the user's wording.",
),
models: promptExplorerInputSchema.shape.models.describe(
"Models to query: chat_gpt, claude, gemini, perplexity. Defaults to ChatGPT only. Supply a list to compare selected models; each uncached model uses credits.",
),
highlightBrand: promptExplorerInputSchema.shape.highlightBrand.describe(
"Optional brand to highlight in answer text and citations. Does not change the prompt sent to the models.",
),
webSearch: promptExplorerInputSchema.shape.webSearch.describe(
"Allow web search, default true. Answers report whether the model actually searched; permission does not guarantee citations.",
),
webSearchCountryCode:
promptExplorerInputSchema.shape.webSearchCountryCode.describe(
"Optional two-letter web-search country code. Omit for no country preference. Gemini does not support country selection; unsupported model/country combinations return per-model errors without a paid call.",
),
});
const [success, failure] = promptExplorerModelResultSchema.options;
const modelOutputSchema = z.discriminatedUnion("status", [
success
.extend({
citations: z.array(success.shape.citations.element.loose()),
})
.loose(),
failure.loose(),
]);
export const explorePromptTool = {
name: "explore_prompt",
config: {
title: "Explore an AI prompt",
description:
"Ask one prompt in ChatGPT, or compare answers from selected AI models. Returns answer text, citations, search queries, brand mentions, and per-model errors. Use for a one-off answer comparison; use AI visibility tracking tools for consumer-site observations and recurring monitoring. Charges actual model usage per uncached model; web search may make one additional paid attempt if the first answer did not search. Cached answers are free for seven days. Requires a paid plan in hosted mode. Does not configure tracking or start project setup.",
inputSchema: inputSchema.shape,
outputSchema: promptExplorerResultSchema
.extend({
results: z.array(modelOutputSchema),
...optionalMetaOutputSchema,
})
.loose(),
annotations: {
readOnlyHint: false,
openWorldHint: true,
destructiveHint: false,
},
},
handler: withMcpProjectAuth(
async (args: z.infer<typeof inputSchema>, context) => {
const { explorePrompt } =
await import("@/server/features/ai-search/services/promptExplorer");
const result = await explorePrompt(args, context.billing);
const meta = buildProjectMeta(
context,
args.projectId,
`/p/${args.projectId}/prompt-explorer`,
{
q: args.prompt,
models: JSON.stringify(args.models),
web: String(args.webSearch),
cc: args.webSearchCountryCode ?? "default",
hb: args.highlightBrand,
},
);
return mcpResponse({
text: `Prompt Explorer: ${result.results.length} model results.\n${meta.url}\n${JSON.stringify(result)}`,
meta,
structuredContent: result,
});
},
),
};