Zhipu web search API alternative in 2026

The call, before and after

# Before: Z.ai Web Search API, $0.01 a call
curl -X POST https://api.z.ai/api/paas/v4/web_search \
  -H "Authorization: Bearer $ZAI_API_KEY" -H "Content-Type: application/json" \
  -d '{"search_engine": "search-prime", "search_query": "GLM-5.3 release date", "count": 10, "search_domain_filter": "www.sohu.com"}'

# After: SpecStream Search API, $0.001 a call with speed=fast
curl -G https://api.specstream.ai/v1/web-search \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" \
  --data-urlencode "q=GLM-5.3 release date site:www.sohu.com" \
  -d num=10 -d speed=fast

search_query becomes q and count becomes num. The domain in search_domain_filter goes into the query as site:, and after: with a date in the query does what search_recency_filter does. Your code reads data.sources[]: title stays, link is url, content is snippet (or excerpt with passages=true), media is site and publish_date is published.

Give GLM SpecStream as a function tool

Z.ai answers OpenAI's SDK at https://api.z.ai/api/paas/v4/, and GLM calls functions you define. Define search_web, run the SpecStream search when GLM asks for it, and send the results back as a tool message. GLM writes its reply from them.

import json, os, requests
from openai import OpenAI

glm = OpenAI(api_key=os.environ["ZAI_API_KEY"], base_url="https://api.z.ai/api/paas/v4/")

tools = [{
    "type": "function",
    "function": {
        "name": "search_web",
        "description": "Search the live web. Returns ranked results with title, URL and snippet.",
        "parameters": {"type": "object", "properties": {"q": {"type": "string"}}, "required": ["q"]},
    },
}]

def search(q):
    data = requests.get(
        "https://api.specstream.ai/v1/web-search",
        params={"q": q, "num": 10, "speed": "fast"},
        headers={"Authorization": f"Bearer {os.environ['SPECSTREAM_API_KEY']}"},
        timeout=30,
    ).json()["data"]
    return json.dumps([{"title": s["title"], "url": s["url"], "snippet": s["snippet"]} for s in data["sources"]])

messages = [{"role": "user", "content": "Who is the CEO of Mistral AI?"}]
while True:
    reply = glm.chat.completions.create(model="glm-5.3", messages=messages, tools=tools).choices[0].message
    if not reply.tool_calls:
        break
    messages.append(reply)
    for call in reply.tool_calls:
        messages.append({"role": "tool", "tool_call_id": call.id,
                         "content": search(json.loads(call.function.arguments)["q"])})

print(reply.content)

In Cursor, Cherry Studio or any MCP client, add https://api.specstream.ai/mcp beside GLM in place of Z.ai's web search server. The MCP guide covers each client, and the Cherry Studio setup covers Cherry Studio.

From $1 per 1,000, a tenth of Z.ai's $10

100,000 searches cost $1,000 on Z.ai at $0.01 a use. On SpecStream they cost $100 with speed=fast and $500 on the default search, which adds the answer and the page it was read on. Each reply carries price_usd, and the account's request log and usage totals show every call and its price.

The most accurate web search API in the world

The SpecStream Search API scores a 97.9 F-score on all 1,000 official SimpleQA Verified questions, with 978 correct, and 64.0 item F1 and 43.4 row F1 on all 200 official WideSearch tasks. Scores and method are at specstream.ai/benchmarks.

Choose which engines your searches reach

A query written in Chinese is searched in Chinese, on the engines that answer Chinese best. GET /v1/search-engines lists every engine with the country it is run from, and a key made with "engine_rule": {"deny": ["CN"]} never reaches an engine run from China. Set zero_retention on a key and no request text and no answer is kept after the reply.

SpecStream

The live web data engine for teams and agents, and the best alternative to the Z.ai Web Search API. The Search API costs $5 per 1,000 searches, or $1 with speed=fast, works with GLM and any other model and scores 97.9 F-score on all 1,000 SimpleQA Verified questions. The Extract API costs $0.80 per 1,000 pages. Deep Search returns every match as structured rows at 5¢, 15¢ or $1 a row.

Bocha Web Search API

A Chinese search API at https://api.bochaai.com/v1/web-search, with count from 1 to 50. SpecStream answers Bocha's own call in Bocha's JSON at $1 per 1,000 when you change the base address. Bocha Web Search API compatible.

OpenAI web search tool

A tool on the OpenAI API at $10 per 1,000 calls, plus search content tokens billed at the model's rates. SpecStream costs from $1 per 1,000 searches with no tokens, for any model. OpenAI web search tool alternative.

Claude web search tool

A server tool on the Claude API at $10 per 1,000 searches, plus standard token costs for the content it brings back. SpecStream returns the results to Claude, GLM or any other model from $1 per 1,000. Claude web search tool alternative.

xAI web search tool

The web_search tool on the xAI API at $5 per 1,000 calls. SpecStream costs from $1 per 1,000 with no tokens billed. xAI web search tool alternative.