Bing Web Search API Alternative in 2026 — fastCRW [Raw Results + Scrape, Self-Host, No Azure Lock-In]

If your Bing Web Search API was retired and you need a replacement, this page is a sourced migration guide. Microsoft's official replacement is answer-only and Azure-locked; fastCRW is a raw-results-plus-scrape API you can self-host. Below is the honest framing of what changed and where each option fits.

Verdict

Bing Web Search API is retired; here is a like-for-like migration target. Microsoft announced the retirement on 2025-05-15 and fully decommissioned the API on 2025-08-11 — no new sign-ups, and existing instances were shut down (per learn.microsoft.com/lifecycle/announcements/bing-search-api-retirement, verified 2026-07).

The official replacement path is Grounding with Bing Search inside Azure AI Agents / Azure AI Foundry. It is deliberately narrower than the API it replaces: it is answer-only — there is no raw SERP endpoint — and it is locked to the Azure agent stack, priced around $14 per 1,000 transactions (per microsoft.com/en-us/bing/apis, verified 2026-07). So teams that relied on the retired API's raw search-results JSON have no like-for-like Microsoft product anymore.

Choose fastCRW when you want raw search results plus page content — for RAG, rank tracking, or agents — the way the old Bing API worked, without Azure lock-in. Choose Grounding with Bing when you are all-in on Azure AI Foundry and want grounded answers rather than raw results.

Why You Need a Replacement

The Bing Web Search API is not deprecated — it is gone. The migration is forced, not optional:

1. Retirement is complete. Announced 2025-05-15, fully decommissioned 2025-08-11: no new sign-ups, existing instances shut down (per learn.microsoft.com/lifecycle/announcements/bing-search-api-retirement, verified 2026-07).
2. The replacement is answer-only. Grounding with Bing Search returns grounded model answers, not a raw SERP result list. There is no drop-in endpoint that returns the same result JSON the retired API did (per learn.microsoft.com/azure/ai-services/agents/how-to/tools/bing-grounding, verified 2026-07).
3. The replacement is Azure-locked. Grounding with Bing lives inside Azure AI Agents / Azure AI Foundry — you adopt the Azure agent stack to use it, which is a hard requirement change for anyone who called the old REST API directly.
4. It costs more. Grounding with Bing is priced around $14 per 1,000 transactions (per microsoft.com/en-us/bing/apis, verified 2026-07); third-party analyses report materially higher effective costs than the retired API depending on tier and usage.

If You're All-In on Azure

Honest version: if your stack is already Azure AI Foundry, Grounding with Bing is the native path — and for some teams the right one.

• It is the first-party option. If you are building on Azure AI Agents and want Bing-backed grounding inside that runtime, Grounding with Bing is the supported, integrated tool.
• Answer-style grounding, not raw results. If your product wants a grounded model answer with citations rather than a result list you re-rank yourself, the answer-only shape is a feature, not a limitation.
• Managed and Azure-billed. If you already run everything through Azure billing and identity, keeping search there avoids adding a vendor.

fastCRW does not replace that Azure-native grounding experience. It replaces the raw search results plus page content the retired REST API gave you.

Head-to-Head

Decision area	fastCRW	Retired Bing Web Search API	Grounding with Bing (replacement)
Availability	Live	Retired — decommissioned 2025-08-11	Live (Azure-locked)
Raw search results	Yes — /v1/search returns raw results	Yes (retired)	No — answer-only
Search + scrape in one API	Yes — /v1/search and /v1/scrape share auth and credits	No — SERP only	No — answer-only
Self-host shape	Single AGPL-3.0 binary, not cloud-locked	SaaS only (retired)	Azure AI Foundry only
MCP	Built-in MCP server in the core binary	No	No first-party MCP
Pricing model	Flat credits per request, plus self-host	Per-transaction (retired)	~$14 / 1,000 transactions (per microsoft.com/en-us/bing/apis, verified 2026-07); third-party analyses report materially higher effective costs
Effective search price	$0.47/1,000 annual ($0.55 monthly; 1 credit/query; derived from PLAN_DISPLAY / /pricing, verified 2026-07) — same credit also scrapes	(retired)	~$14 per 1,000 transactions (per microsoft.com/en-us/bing/apis, verified 2026-07)
Concurrency	Up to 150 concurrent requests on the Scale plan (derived from PLAN_DISPLAY / fastCRW /pricing, verified 2026-07)	(retired)	Azure-metered per transaction
Best fit	RAG, rank tracking, agents wanting raw results + content	(gone)	Azure-native grounded answers

Availability rows cite Microsoft's docs; the $14/1,000 rate is per microsoft.com/en-us/bing/apis (verified 2026-07); higher effective-cost estimates are third-party analyses.

Where fastCRW Wins

• A raw-results path that still exists. /v1/search returns raw search results plus optional clean markdown in the same call — the like-for-like "raw results" the retired Bing API offered, and exactly what answer-only Grounding with Bing does not give you.
• Search + scrape in one API. One credential and one credit pool cover both the search step and reading the pages behind the results.
• AGPL-3.0 single-binary self-host — a single small Rust binary that runs on any Linux box, so search is not tied to Azure or any managed cloud.
• Built-in MCP so any MCP-aware agent gets crw_search and crw_scrape as native tools.
• Flat, predictable credits — search is 1 credit per query, with optional content scraping in the same call — instead of per-transaction Azure billing.

fastCRW search averaged 880 ms over a 100-query benchmark (benchmarks/triple-bench.ts, concurrent vs Firecrawl + Tavily); 73/100 latency wins; 100% success. This is a search benchmark only — it does not measure scrape performance.

Cost and Migration Note

The retired API billed per transaction directly. Grounding with Bing is priced around $14 per 1,000 transactions (per microsoft.com/en-us/bing/apis, verified 2026-07), and third-party analyses report materially higher effective costs than the retired API depending on tier and usage. It also carries the implicit cost of adopting the Azure agent stack.

fastCRW uses flat credits per request — search is 1 credit per query plus optional per-result content scraping. fastCRW is the cheapest real-time web search API per 1,000 requests — about $0.47 per 1,000 searches on an annual plan (~$0.55 monthly; 1 credit/query; derived from PLAN_DISPLAY / /pricing, verified 2026-07), responding in ~880 ms (100-query benchmark; search benchmark only — does not measure scrape performance). The same credit also scrapes the full page and exposes MCP — a fraction of Grounding with Bing's ~$14 per 1,000 transactions (per microsoft.com/en-us/bing/apis, verified 2026-07), with up to 150 concurrent requests on Scale and no Azure lock-in. It is also the only option here you can self-host — an AGPL-3.0 single binary — so teams that cannot route queries through a third-party API pay for infrastructure, not per search. See /pricing for current rates.

Migration Path

fastCRW exposes a Firecrawl-compatible API surface, so replacing a retired Bing Web Search API call is a client swap that keeps the raw-results shape. Example in Python:

import os
import httpx

FASTCRW_BASE = "https://api.fastcrw.com/v1"
HEADERS = {"Authorization": f"Bearer {os.environ['FASTCRW_API_KEY']}"}

async def search_and_read(query: str, top_k: int = 5) -> list[dict]:
    async with httpx.AsyncClient(timeout=30.0) as client:
        # 1. Search step — replaces the retired Bing Web Search API,
        #    and returns raw results (not an answer-only response).
        serp = await client.post(
            f"{FASTCRW_BASE}/search",
            headers=HEADERS,
            json={"query": query, "limit": top_k},
        )
        serp.raise_for_status()
        results = serp.json()["data"]

        # 2. Read step — same auth, same vendor, no Azure stack required.
        pages: list[dict] = []
        for hit in results:
            page = await client.post(
                f"{FASTCRW_BASE}/scrape",
                headers=HEADERS,
                json={"url": hit["url"], "formats": ["markdown"]},
            )
            page.raise_for_status()
            pages.append(page.json()["data"])
        return pages

The same shape works in TypeScript with the crw-ts SDK. The structural change from a retired-Bing-API stack is pointing your existing search-results consumer at /v1/search — the raw-results contract is preserved, unlike the answer-only Grounding replacement.

Recommended Evaluation Flow

1. Run your real queries in the playground and confirm the raw search results plus rendered page output match what you consumed from the retired Bing API.
2. Read the 1,000-URL benchmark for coverage and latency distribution in context.
3. Review the benchmark methodology so you can reproduce the workload on your own URLs.
4. Compare against the search docs and scrape docs for endpoint shape.
5. Wire the MCP server into Claude Desktop or Cursor and let the agent search and scrape as native tools.
6. If you are committed to Azure AI Foundry and want grounded answers rather than raw results, keep Grounding with Bing for that surface.

The decision is workload-specific. Bing Web Search API is gone, and Microsoft's replacement is answer-only, Azure-locked, and costlier. fastCRW is the stronger migration target when you want raw search results plus content — for RAG, rank tracking, or agents — with an MCP-ready path and an AGPL-3.0 self-host option.