Workbench developer docs中文

Serper

Base URL

https://desktop-api.wbench.io/v1/proxy/serper/

Use https://desktop-api.wbench.io/v1/proxy/serper/ where Serper's documentation uses https://google.serper.dev/.

It's Serper's own API

Everything after the base URL is Serper's API, exactly as Serper documents it: the same paths, methods, query parameters and JSON bodies. Only the address and the credential change: you send your Workbench token, and Workbench adds your company's Serper key on the server. You never see that key.

Notes

Examples

Set WB_TOKEN first: export WB_TOKEN=$(wb token).

wb

wb call serper search --data '{"q": "MMBT3904 datasheet"}'

curl

curl "https://desktop-api.wbench.io/v1/proxy/serper/search" \
  -H "Authorization: Bearer $(wb token)" \
  -H "Content-Type: application/json" \
  -d '{"q": "MMBT3904 datasheet"}'

Python

Serper has no official SDK for Python or TypeScript, so these examples use plain HTTP.

import os
import requests

headers = {"Authorization": f"Bearer {os.environ['WB_TOKEN']}"}
r = requests.post(
    "https://desktop-api.wbench.io/v1/proxy/serper/search",
    headers=headers,
    json={"q": "MMBT3904 datasheet"},
    timeout=60,
)
r.raise_for_status()
print(r.json())

TypeScript

const r = await fetch("https://desktop-api.wbench.io/v1/proxy/serper/search", {
  method: "POST",
  headers: { Authorization: `Bearer ${process.env.WB_TOKEN}`, "Content-Type": "application/json" },
  body: JSON.stringify({"q": "MMBT3904 datasheet"}),
});
if (!r.ok) throw new Error(`${r.status}: ${(await r.json()).message}`);
console.log(await r.json());

See Limits and errors for what a refusal looks like.