# Grokipedia Search API

> Search Results from grokipedia.com with description and title fields, plus search inputs.

- Page: https://cleanedweb.com/catalog/grokipedia-search-api/
- Contract (JSON): https://cleanedweb.com/catalog/grokipedia-search-api/contract.json
- Source: grokipedia.com (https://grokipedia.com/)
- Category: News & content
- Updated: 2026-11-18

A run reads until the source ends or your `max_items`, `timeout_seconds` or budget stops it.

## Key fields

| Field | Type | Sample value |
| --- | --- | --- |
| `title` | string | `YouTube` |
| `description` | string | `YouTube is an online video-sharing and social media platform owned by Alphabet …` |
| `url` | string | `https://grokipedia.com/page/YouTube` |
| `search_result_link` | boolean | `true` |
| `slug` | string | `YouTube` |
| `owner` | string | `Google` |
| `status` | string | `Active` |
| `references` | string | `352 sources` |

Sample values come from one saved record, not a live run. Every field with its type, and the schemas, are in https://cleanedweb.com/catalog/grokipedia-search-api/contract.json

## Inputs

- `search` (string \| array of string): Optional. Text for the site's search box. Leave it out to use the default: youtube.

`enrich_records` is on by default and adds the fields that only appear on each record's detail page. Send `false` to read list fields only. See https://cleanedweb.com/docs/run-api/

## Request

`POST https://api.cleanedweb.com/v1/catalog/public/apis/generated_4e688315f9e70c47abcc1bbd/endpoints/generated_4e688315f9e70c47abcc1bbd_search/run`

```bash
curl -X POST https://api.cleanedweb.com/v1/catalog/public/apis/generated_4e688315f9e70c47abcc1bbd/endpoints/generated_4e688315f9e70c47abcc1bbd_search/run \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Idempotency-Key: UNIQUE_REQUEST_ID" \
  -H "Content-Type: application/json" \
  -d '{
  "input": {},
  "version": 1,
  "max_items": 25
}'
```

Send an `Idempotency-Key` header with every run and save it before you send. Reuse it only to retry the same request. If the run does not end within `wait_seconds`, read the result with the same API key and key:

`GET https://api.cleanedweb.com/v1/catalog/public/apis/generated_4e688315f9e70c47abcc1bbd/endpoints/generated_4e688315f9e70c47abcc1bbd_search/result`

## Versions

Current: `1.0`. Pin a major with "version": N in the run body (an integer). Leave it out and the call runs on the latest major. A sunset major runs on the latest major and the response says rerouted. An unknown major answers 404 catalog_version_not_found and lists the available majors.

- v1: current, newest minor `1.0`

### Changelog

- `1.0` (major) 2026-11-17

The same changelog, as data, is in https://cleanedweb.com/catalog/grokipedia-search-api/contract.json

## Use it from an AI agent

MCP server: https://api.cleanedweb.com/mcp?profile=catalog (Streamable HTTP). Sign in with a CleanedWeb account when the client asks (OAuth), or send an API key with the `read` and `run` scopes as `Authorization: Bearer`. Call `get_api` with `api_id` `generated_4e688315f9e70c47abcc1bbd`, then `run_api` with the `run_with` arguments it returns. See https://cleanedweb.com/docs/mcp/

## Run it

Reading the catalog needs no credential. Running needs a CleanedWeb account and an API key with the `run` scope; the first 5,000 records are free, then $1 per 1,000 records. See https://cleanedweb.com/docs/run-catalog-api/

CleanedWeb is not affiliated with Grokipedia.
