Semantic Kernel web search after Bing and Google

One option per connector

BingTextSearch uses options?.Endpoint when you set it and Bing's retired address when you do not. Before and after:

using Microsoft.SemanticKernel.Plugins.Web.Bing;

var textSearch = new BingTextSearch(apiKey: "<your Bing key>");

using Microsoft.SemanticKernel.Plugins.Web.Bing;

var textSearch = new BingTextSearch(
    apiKey: Environment.GetEnvironmentVariable("SPECSTREAM_API_KEY"),
    options: new BingTextSearchOptions
    {
        Endpoint = new Uri("https://api.specstream.ai/v7.0/search"),
    });

GoogleTextSearch hands its BaseClientService.Initializer to Google's CustomSearchAPIService, so BaseUri decides the host. Pass any engine ID: the constructor refuses an empty one, and SpecStream accepts cx and does not use it. Before and after:

using Microsoft.SemanticKernel.Plugins.Web.Google;

var textSearch = new GoogleTextSearch(
    searchEngineId: "<your engine ID>",
    apiKey: "<your Google key>");

using Google.Apis.Services;
using Microsoft.SemanticKernel.Plugins.Web.Google;

var textSearch = new GoogleTextSearch(
    initializer: new BaseClientService.Initializer
    {
        ApiKey = Environment.GetEnvironmentVariable("SPECSTREAM_API_KEY"),
        BaseUri = "https://api.specstream.ai/",
    },
    searchEngineId: "any");

In Java, Semantic Kernel's BingConnector sample reads bing.search.url, then BING_SEARCH_URL, then falls back to Bing's address:

# Before: the sample's BingConnector calls https://api.bing.microsoft.com/v7.0/search
export BING_API_KEY=<your Bing key>

# After
export BING_SEARCH_URL=https://api.specstream.ai/v7.0/search
export BING_API_KEY=$SPECSTREAM_API_KEY
# or start the JVM with -Dbing.search.url=https://api.specstream.ai/v7.0/search

The same request and the same JSON

BingTextSearch sends GET ?q=...&count=...&offset=... with the key in Ocp-Apim-Subscription-Key, and GoogleTextSearch sends Google's customsearch/v1 call with key and cx. Both ask for 3 results unless you set Top. SpecStream answers each in its own shape: webPages.value[] for Bing, items[] and searchInformation.totalResults for Google.

The connectors' own code read these replies on production on November 10, 2026: BingTextSearchOptions.Endpoint and GoogleTextSearch with BaseUri in Microsoft.SemanticKernel.Plugins.Web 1.82.0-alpha, and the Java sample's parser on BING_SEARCH_URL. The same Bing call from Python:

import os, requests

reply = requests.get(
    "https://api.specstream.ai/v7.0/search",
    headers={"Ocp-Apim-Subscription-Key": os.environ["SPECSTREAM_API_KEY"]},
    params={"q": "open source vector databases", "count": 3, "offset": 0},
).json()

for page in reply["webPages"]["value"]:
    print(page["name"], page["url"], page["snippet"])

Every parameter and field of each route is in the Bing Web Search v7 guide and the Google Custom Search guide.

$5 per 1,000 searches, with no daily limit

A search of up to 10 results costs $0.005. There is no monthly plan and no daily limit. Credit you buy never expires, and a new account starts with $10 of credit, which covers 2,000 searches. A call stopped by a 429 charges nothing.

The most accurate web search API, in about one second

On SimpleQA Verified the SpecStream Search API scores 97.9 F-score on all 1,000 official questions, 978 of them correct. On WideSearch it scores 64.0 item F1 and 43.4 row F1 on all 200 official tasks. The Extract API scores 0.908 F1 on WCXB. Every score and its method is on the benchmarks page.

Every drop-in call runs the fast search. Measured on production on November 9, 2026, its median was 0.82 s at 1 search a second and 1.07 s at 20 a second, the 95th percentile stayed between 1.09 and 1.73 s, and it answered 718 of 720. One account can send up to 26,666 searches a minute.

You decide what is kept, on each key

Create the key with zero_retention and a search made with it keeps no request text and no answer after the reply. spend_cap_cents is the most the key can spend.

curl -X POST https://api.specstream.ai/v1/account/keys \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
  -d '{"name": "production", "scopes": ["account:read", "finds:spend"], "spend_cap_cents": 500000, "zero_retention": true}'

Search, pages and rows on one key

The same key reads up to 20 pages or documents a call with the Extract API at $0.80 per 1,000 pages, and a page with no text is not charged. Deep Search returns every match for a request in plain words as structured rows, at 5¢, 15¢ or $1 a row, with a source URL on every value. The most a run can cost is approved before it starts.

Google Custom Search, until January 1, 2027

Existing customers can keep GoogleTextSearch on Google's address until Google ends the API: $5 per 1,000 queries after 100 free a day, up to 10,000 a day. A new project gets no key. SpecStream costs the same $5 per 1,000, with no daily limit, behind one BaseUri.

Grounding with Bing Search

Microsoft's named replacement for the Bing Search APIs, at $14 per 1,000 transactions, inside Azure AI Agents. It returns an answer with citations, and developers get no access to the raw results, so BingTextSearch has nothing to read. SpecStream returns Bing's webPages.value[] at $5 per 1,000.

Brave Search API

Semantic Kernel's BraveTextSearch calls Brave at $5 per 1,000 requests, with $5 in free credits a month and up to 50 queries a second. It is a different connector with a different reply. SpecStream keeps the Bing or Google connector you have, at the same $5 per 1,000 and 97.9 on SimpleQA Verified.

Tavily

Semantic Kernel's TavilyTextSearch calls https://api.tavily.com/search, billed in credits at $0.008 each on pay as you go, 1 credit a basic search. On a monthly plan, requests stop when the month's credits run out. SpecStream charges $0.005 a search and its credit never expires.