Google Search API comparison
FlashData vs SerpApi
Pay as you go, or save with a monthly plan.
Move your search workflow into one API for Google results, YouTube data and media delivery. Compare the cost and code, then run your first task.
01 · At a glance
Feature-by-feature comparison
Compare the cost to start, the search fields your app needs, and how an agent submits and retrieves work.
| Feature | FlashData | SerpAPI |
|---|---|---|
| Lower cost to start | ||
| Google Search · entry paid price per 1,000 | $1.00Pay as you go | $25.00Starter · full monthly allowance |
| Starting paid commitment | $10Credit top-up · no monthly fee | $25.00/mo1,000 searches included |
| Initial free Google Search allowance | Up to 2,000 searchesOne-time signup allowance | 250 searches / monthRecurring Free plan |
| Search coverage and usable evidence | ||
| Web, images, news, shopping, videos and places | YesSix Google sources | YesGoogle engines and search types |
| Country, language and local targeting | gl, hl and locationSettings in params | gl, hl and locationAlso supports encoded uule targeting |
| Date filters for current research | Yestbs · past day, week or other supported filters | Yestbs · Google time filters |
| Retrieve the next result page | Page numberparams.page · one page per query | Result offsetstart · 0, 10, 20 for ten-result pages |
| Titles, source URLs and snippets | Organic result arraytitle, link, snippet and position | Organic result arraytitle, link, snippet and position |
| Related questions and rich results | YesPeople Also Ask, related searches, answer box and knowledge graph when returned | YesRelated questions, related searches, answer box and knowledge graph when returned |
| Choose a device for rank tracking | No published device selectorOrganic position in the returned search | Desktop, tablet or mobileSelect device in the request |
| Batch work and saved results | ||
| Realtime and persisted async searches | YesReceive results now, or save a job ID and retrieve later | Yesasync=true, then Search Archive retrieval |
| Queries in one JSON Batch submission | 100 on pay as you goUp to 500 by plan | One search per requestSubmit multiple requests from your client |
| Markdown for agent context | YesAPI response option and CLI output | YesCLI Markdown output |
| Saved search result availability | 24 hours after completionAsync result JSON; save a durable copy | 31 daysSearch Archive |
| Agent access and connected workflows | ||
| Try your own agent task before owner signup | YesUp to 50 Google searches · 72 hours | Preset CLI demo queriesAn account key is needed for custom searches |
| Hosted MCP authentication | OAuth or API keySign in and approve the connection | API keyConfigure the key in your client |
| Standalone CLI · no runtime to install | YesLinux, macOS and Windows | YesNative CLI executables |
| Owner claim and credential handoff | YesApprove access; keep the same CLI profile | Configure an account API key |
| YouTube video / audio delivery to your storage | YesYouTube Download API | Not listedIn the API catalog |
The entry unit price compares FlashData pay as you go with SerpApi Starter at its full monthly allowance; other plans have different rates. FlashData signup and agent trials are separate, subject to eligibility. Signup counts use the full trial balance for search. Not listed refers to the published API catalog.
SerpAPI alternatives for Google search JSON
Compare plans and monthly usage →See the migration example →
Pay for the workload
Use prepaid credits without a subscription, or choose included credits and higher capacity for steady usage.
Google Search pricing →Continue into YouTube
Retrieve metadata and captions, then deliver video or audio to your storage through the same account.
Build a research workflow →Let the agent try a task
Eligible agents can start with temporary access. The owner then claims the connection and approves ongoing use.
See the agent flow →02 · Compare costs
Price your monthly searches.
Count successful, billable queries across your search pages and scheduled runs. The same result count can require different numbers of queries in your application.
FlashData Google Search plans
| Plan | Per 1,000 requests | Monthly fee | Included credits | Google Search requests |
|---|---|---|---|---|
| Pay As You Go | $1.00 | $0 · no monthly fee | None | Top up as needed |
| Micro | $0.80 | $49.00 | 1,000 ($50.00 of usage) | ≈ 62,500 |
| Starter | $0.65 | $199.00 | 4,000 ($200.00 of usage) | ≈ 307,692 |
| Pro | $0.60 | $499.00 | 10,000 ($500.00 of usage) | ≈ 833,333 |
| Scale | $0.50 | $999.00 | 20,000 ($1,000.00 of usage) | ≈ 2,000,000 |
| Business | $0.40 | $1,999.00 | 40,000 ($2,000.00 of usage) | ≈ 5,000,000 |
Included credits renew each billing period. Request counts assume all included credits go to Google Search; Google Search, YouTube Data and downloads share the same balance.
View full Google Search pricing →Estimate your monthly bill
Each page of results is another query.
FlashData
Pay As You Go
$1.00per 1,000 searches · plan usage rate
- Monthly fee
- $0.00
- Included requests / month
- 0
- Monthly bill at 10,000 searches
- $10.00
- Effective cost / 1,000 at this volume
- $1.00
Prepaid usage. Top up as needed.
25 concurrent requests · 100 queries / Batch
SerpApi
Production
$10.00per 1,000 searches · at full plan allowance
- Monthly fee
- $150.00
- Included requests / month
- 15,000
- Monthly bill at 10,000 searches
- $150.00
- Effective cost / 1,000 at this volume
- $15.00
5,000 searches unused at this volume.
3,000 searches / hour
USD. FlashData shows usage plus any selected plan fee; credit purchases start at $10 and can leave a balance for later. Trial credits are excluded.
How the estimate works
Choose a plan independently for each provider. FlashData’s unit price is the selected plan’s usage rate; included requests assume all shared credits go to Google Search. SerpApi’s unit price is the monthly plan price divided by its full search allowance, multiplied by 1,000. Auto-select chooses the smallest listed monthly Best Effort plan covering the volume, including Free. A manually selected SerpApi plan stays selected if its allowance is too small; the calculator shows the shortfall instead of a workload bill or effective rate. Effective cost per 1,000 searches is the monthly bill divided by your entered volume, multiplied by 1,000; it is not calculated at zero searches. FlashData: monthly fee plus usage above included credits; pay as you go has no monthly fee. Self-service credit top-ups start at $10. Purchases are prepaid and can exceed the modeled usage, leaving a balance for later. No trial credits, tax, annual discounts, speed add-ons, early renewals, custom contracts or other API usage are included. Concurrency and hourly throughput are different limits.
At small volumes, SerpApi’s recurring Free plan can cost less than metered FlashData usage. Select “Lowest modeled cost” to include FlashData subscriptions and their shared credits.
| Plan | Per 1,000 searches | Searches / month | Price / month | Searches / hour |
|---|---|---|---|---|
| Free | Free | 250 | $0.00 | 50 |
| Starter | $25.00 | 1,000 | $25.00 | 200 |
| Developer | $15.00 | 5,000 | $75.00 | 1,000 |
| Production | $10.00 | 15,000 | $150.00 | 3,000 |
| Big Data | $9.17 | 30,000 | $275.00 | 6,000 |
| Searcher | $7.25 | 100,000 | $725.00 | 20,000 |
| Volume | $5.90 | 250,000 | $1,475.00 | 50,000 |
| Infrastructure | $5.50 | 500,000 | $2,750.00 | 100,000 |
| Cloud 1M | $3.75 | 1,000,000 | $3,750.00 | 110,000 |
FlashData: start with trial credits
Eligible new accounts can try up to 2,000 Google searches after signup and email verification, using the full shared trial balance for search. No card required. This is a signup allowance.
Signup trial details →SerpApi: a recurring Free plan
250 searches per month, with 50 searches per hour. The calculator includes this plan. FlashData’s separate temporary agent trial is described below.
Agent trial and claiming →03 · Migrate a search
Keep your task. Change the provider call.
This example retrieves US English Google results for openai. Both versions return the same application fields: position, title, URL and snippet.
1 Create a FlashData key
Create or sign in to your account, complete the account setup, and open API Keys. Enable google_search and check your available credits. Keep the key in your server’s environment or secret store.
Using an AI agent? Start through its saved profile or eligible trial.
2 Run one request
Python 3, standard library only. Save either complete example in your task folder using the filename shown, then run it. Each script reads its environment variable or prompts privately for a key, performs one search and prints normalized results.
SerpApi
import getpass, json, os
from urllib.parse import urlencode
from urllib.request import Request, urlopen
from urllib.error import HTTPError, URLError
key = os.environ.get("SERPAPI_API_KEY")
if not key:
key = getpass.getpass("SerpApi key: ")
params = {"engine": "google", "q": "openai",
"gl": "us", "hl": "en", "api_key": key}
request = Request("https://serpapi.com/search.json?" + urlencode(params))
try:
with urlopen(request, timeout=120) as response:
body = json.load(response)
except HTTPError as error:
raise SystemExit(f"HTTP {error.code}: check input and access.") from None
except (URLError, TimeoutError):
raise SystemExit("No response; check state before retrying.") from None
if body.get("search_metadata", {}).get("status") != "Success":
raise SystemExit("Search is not successful; check its status.")
items = body.get("organic_results", [])
print(json.dumps([
{"position": item.get("position"), "title": item.get("title"),
"url": item.get("link"), "snippet": item.get("snippet")}
for item in items
], indent=2, ensure_ascii=False))Run: python3 serpapi-search.py
FlashData
import getpass, json, os
from urllib.request import Request, urlopen
from urllib.error import HTTPError, URLError
key = os.environ.get("FLASHDATA_API_KEY")
if not key:
key = getpass.getpass("FlashData key: ")
payload = {"source": "google_search", "query": "openai",
"params": {"gl": "us", "hl": "en", "num": 10}}
request = Request("https://data.flashdata.dev/v1/queries/realtime",
data=json.dumps(payload).encode(),
headers={"X-API-Key": key, "Content-Type": "application/json"})
try:
with urlopen(request, timeout=120) as response:
body = json.load(response)
except HTTPError as error:
raise SystemExit(f"HTTP {error.code}: check input and access.") from None
except (URLError, TimeoutError):
raise SystemExit("No response; check state before retrying.") from None
if body.get("status") != "done":
raise SystemExit("Search is not complete; check its status.")
items = (body.get("results") or [{}])[0].get("organic", [])
print(json.dumps([
{"position": item.get("position"), "title": item.get("title"),
"url": item.get("link"), "snippet": item.get("snippet")}
for item in items
], indent=2, ensure_ascii=False))Run: python3 flashdata-search.py
3 Read the normalized result
Expect an array of source records like this illustration. The examples check completion before extracting results, and map each provider’s link to your application’s url.
Titles, order and snippets are live data. An empty array can be valid; do not treat it as a reason to repeat the request automatically.
Full FlashData request and response →[
{
"position": 1,
"title": "OpenAI | Research & Deployment",
"url": "https://openai.com/",
"snippet": "..."
}
]Request and response mapping
| Change | SerpApi | FlashData | Migration note |
|---|---|---|---|
| HTTP request | GET /search.json | POST /v1/queries/realtime | Send a JSON body to data.flashdata.dev. |
| Authentication | api_key | X-API-Key header | Create a FlashData key with google_search permission. |
| Search | engine=google · q | source: google_search · query | Keep your search expression, such as site:example.com. |
| Country and language | gl · hl | params.gl · params.hl | Set both explicitly for your target market. |
| Location and dates | location · tbs | params.location · params.tbs | Supported inputs; location labels and filter behavior must suit your query. |
| Pagination | start=0, 10, 20… | params.page=1, 2, 3… | Use page numbering. Request num=10 for this ten-result-page example; bill each page separately. |
| Organic results | organic_results[] | results[0].organic[] | title, link, snippet and position remain useful application fields. |
| Related results | related_questions · related_searches | results[0].peopleAlsoAsk · results[0].relatedSearches | Map fields explicitly; optional sections can be absent. |
Check advanced parameters before switching. FlashData’s Google Search schema does not accept SerpApi’s uule, device, google_domain, safe, udm or no_cache. Use the dedicated FlashData source for another search type; do not pass SerpApi’s entire parameter object through.
Async jobs, retries and saved results
| Behavior | SerpApi | FlashData | What to change |
|---|---|---|---|
| Completion | search_metadata.status: Success | status: done | Read HTTP status and the body status. An empty result list can still be a completed, billable search. |
| Async retrieval | async=true → search_metadata.id → Search Archive | POST /v1/queries → id → /v2/jobs/{id} | Save the job ID immediately. Retrieve the finished output through the job result endpoint. |
| Result retention | Archive: up to 31 days after completion | Async result JSON: 24 hours after completion | Export FlashData results to your own storage before expiry. |
| Repeated work | Exact-parameter cache: 1 hour; no_cache controls it | CLI run ID reuses completed local results | These are different mechanisms. Keep the CLI profile, input, account and state directory; use a new run ID for fresh data. |
| Quota or capacity errors | 429 can mean hourly throughput or exhausted searches | Inspect the structured error code and account balance/limits | Do not retry all 429s blindly. Back off for capacity; resolve exhausted credits or revoked access. |
FlashData job reads use https://api.flashdata.dev/v2/jobs/{id}; completed results use /v2/jobs/{id}/results. Follow the result redirect without forwarding your API key to storage.
Async Python and cURL examples →Choose Batch or Bulk →Errors and recovery →
04 · For AI agents
From one search to a recurring task.
Both services have MCP and CLI tools. FlashData adds a temporary-account claim flow: test an eligible task, then connect the owner’s account for ongoing use.
- Reuse or start access
Check saved credentials first. Install the CLI or connect MCP; request only the permissions the task needs.
- Complete the first task
Inspect costs, run one query, and save the result and receipt. Keep the same profile throughout.
- Claim with the owner
The owner signs in or signs up, verifies email and approves permissions and shared-balance use. The client saves its new credential.
- Schedule and recover
Reuse owner access. Give each scheduled occurrence a new run ID; retry that occurrence with the same ID and saved state.
Current agent trial: 1 credit, up to 50 Google searches, 72 hours. Eligibility and registration limits apply. This is a starting allowance; owner usage draws from the owner’s balance.
Give your agent the migration task
Read https://flashdata.dev/compare/serpapi/reference.md and https://flashdata.dev/auth.md.
Migrate only my Google organic-search task. Inspect the existing SerpApi parameters
and consumed response fields first; report unsupported requirements before changing them.
Reuse existing FlashData access. If none exists, check eligibility before starting one trial.
Keep credentials in the host secret store, outside chat and source control.
Run one search for "openai" with gl=us and hl=en. Return title, URL, snippet and position.
Save the raw result and report the actual credit usage. Do not purchase credits.
For recurring use, ask the owner to claim access and approve permissions and shared-balance use.
Keep the same profile and state directory. Stop on revoked access or insufficient credits.Run through an existing CLI profile
After CLI installation and authentication, these commands check access, validate the input, then run one search. All use the default profile.
flashdata auth status --json --non-interactive &&
flashdata search "openai" --gl us --hl en --num 10 \
--dry-run --json --non-interactive &&
flashdata search "openai" --gl us --hl en --num 10 \
--run-id serpapi-migration-first --json --non-interactiveStop on missing or revoked access. Claiming does not purchase credits or create a separate spending cap. A repeat of this completed run reuses its local result; a new scheduled occurrence needs a new ID.
Codex setup →OpenClaw setup →Trial and claiming →Recurring tasks and revocation →
05 · Choose your fit
Move the workflows that fit.
Choose FlashData when…
- You need structured Google results for an application, research workflow or scheduled agent task, with pay-as-you-go billing.
- Your workflow continues into YouTube metadata, captions or delivery of media to your own storage.
- You want an agent to try a task, let its owner claim access, and keep using the same saved profile for recurring work.
Keep SerpApi when…
- You need Bing, Amazon, Google Flights, Google Maps reviews or another engine outside FlashData’s source catalog.
- Your application depends on SerpApi-specific SERP sections, raw HTML, uule, device targeting or dedicated AI Overview / AI Mode endpoints.
- You want SerpApi’s recurring free allowance or its longer search archive, and those match your workload.
You can use both: put a provider adapter behind your search function and migrate Google organic-search calls independently of other engines. FlashData returns search results; it does not fetch the full text of every linked web page.
06 · Questions before switching
The details worth settling first.
Can I just change the SerpApi base URL?
No. Change the method, authentication, request body and response reader. FlashData SDKs and MCP tools also have their own schemas. The example above keeps the application output consistent while changing the provider call.
Does one search mean one returned result?
No. A search is a query request, not each organic item. Pagination adds requests, and every query in a FlashData Batch is metered separately. The calculator uses billable queries, not result rows or video-download bytes.
Can I replace every SerpApi engine?
Use FlashData’s source catalog to choose the matching task. Google web search is google_search; images and news use other sources. Specialized SerpApi engines and fields do not automatically map. You can keep those calls on SerpApi and migrate only the tasks FlashData covers.
Do failed or empty searches cost credits?
FlashData does not charge requests marked failed; deductions for failed requests are returned. SerpApi excludes failed and cached searches from its allowance. An empty search can still be successful on either service. A lost response is not proof of failure: recover the saved job or CLI run before submitting again.
Will my agent’s free credits renew every month?
The FlashData temporary trial is an eligible agent’s starting allowance with an expiry, not a recurring grant. After claiming, requests use the owner’s normal balance. Unused temporary credits do not transfer; normal signup offers are handled independently. SerpApi’s Free plan is a monthly allowance.
Can I use FlashData without moving the rest of my application?
Yes. Put the FlashData call behind your existing search function, normalize the result fields, and keep provider-specific features separate. Start with one Google Search workflow. Existing SerpApi requests, keys and billing stay with SerpApi.
Start with one query
Put FlashData behind your next search.
Try the request above, or give your agent the migration prompt.