--- title: "Cheapest Web Search APIs (Aug 2026): the real cost per 1,000 queries" dek: "The cheapest web search APIs in August 2026, ranked by real cost per 1,000 queries. Keiro ~$1/1k search and ~$3/1k all-in via /search/content, Serper for raw SERP, Exa the biggest free tier." category: "comparisons" tags: [pricing, search-api, comparison, rag] author: "Dave" published: 2026-08-15T12:00:00+00:00 updated: 2026-08-15T12:00:00+00:00 url: https://keirolabs.cloud/blogs/comparisons/cheapest-web-search-apis-aug-2026 ---
$1Keiro / 1k search
$3Keiro RAG / 1k all-in
$0.30Serper / 1k at 12.5M credits
9crFirecrawl page, JSON+Enhanced
20kExa free searches / month
TL;DR · what actually costs what, August 2026
- **Cheapest full-featured (search + clean text + optional embeddings in one call):** Keiro, ~$1/1k for search, ~$3/1k all-in for `/search/content` (3 credits). 500 free credits/mo, no card, commercial use allowed. - **Cheapest raw Google SERP:** Serper, from $1/1k down to $0.30/1k at 12.5M credits. No content, no answers, you build extraction. - **Biggest free tier:** Exa, roughly 20,000 free search requests/mo. Search only; contents, answers, and deep search cost credits. - **The unit trap:** "per 1k" means per query, per search, per credit, or per request-plus-token. SerpAPI at from $15/1k, Tavily at $5-8/1k, Firecrawl at $1.66/1k for a search, convert to your real call shape before you compare. - **The spread at scale:** at 1M queries/mo, Keiro is about $1,000 and SerpAPI is $9,170 to $15,000. Same volume, 9x to 15x before hidden costs. - **Where Keiro loses:** raw SERP at extreme volume (Serper), the biggest free tier (Exa), an independent privacy index (Brave), and the full crawl suite (Firecrawl).
## What does "per 1,000 queries" cost in August 2026? The first table is the one every comparison leads with: the headline price per 1,000 searches, normalized to the cheapest tier each provider will sell you today. I pulled every number below from the provider's own pricing page in August 2026. The catch, which the credit-math section gets into, is that "per 1k" does not mean the same operation at every vendor. | Provider | $/1k (entry) | Unit | Free tier | What you get | | --- | --- | --- | --- | --- | | **Serper** | from $1.00 ([pricing](https://serper.dev)) | per search | 2,500 one-time | Raw Google SERP, no content | | **Keiro** | $1.00 ([pricing](/pricing)) | per query | 500/mo | Ranked results, content bundled (3cr) | | **Firecrawl** | $1.66 ([pricing](https://www.firecrawl.dev/pricing)) | per credit | 1,000 cr/mo | Search 2cr/10 results, scrape 1cr/page | | **Brave** | $5.00 ([pricing](https://brave.com/search/api/)) | per 1k req | none (removed Feb 2026) | Independent index, LLM Context included | | **Perplexity** | ~$5 + tokens ([pricing](https://docs.perplexity.ai/docs/getting-started/pricing)) | per req + token | none | Synthesized cited answer | | **Tavily** | $5.00-8.00 ([pricing](https://tavily.com/pricing)) | per credit | 1,000/mo | Inline content, $5 on Growth / $8 PAYG | | **Exa** | $7.00 ([pricing](https://exa.ai/pricing)) | per 1k search | ~20,000/mo | Neural search, contents $1/1k pages | | **SerpAPI** | from $15.00 ([pricing](https://serpapi.com/pricing)) | per search | 250/mo | 40+ engines, metadata only | Two providers sit at the $1 floor for the headline: Serper and Keiro. **They are not the same product at the same price.** Serper returns raw Google SERP, titles, URLs, snippets, and nothing else. Keiro's 1-credit search returns ranked results, and its 3-credit `/search/content` call returns ranked results plus clean markdown plus optional embeddings in one call. The same $1/1k buys one thing at Serper and three things at Keiro.
Horizontal bar ranking of cheapest web search APIs by price per 1,000 queries
The headline per-1k ranking, August 2026. Serper and Keiro at the $1 floor, SerpAPI 15x above it. The headline is only comparable once you convert units.
The middle of the table is where the unit trap lives. **Firecrawl at $1.66/1k is not cheaper than Brave at $5/1k just because the number is smaller, the units differ.** Firecrawl bills per credit, and a search call costs 2 credits per 10 results, rounded up. Tavily ranges from $5/1k on the Growth plan (100,000 credits for $500/mo, [docs](https://docs.tavily.com/documentation/api-credits)) to $8/1k on pay-as-you-go at $0.008/credit. Perplexity's $5/1k is a per-request fee before tokens. SerpAPI's $15/1k buys metadata only at the Developer tier, and its cheapest plan is actually $25/1k at 1,000 searches for $25/mo. The market context matters too. **Google's own Custom Search JSON API, the "official" option, is closed to new customers and shuts down entirely on January 1, 2027** ([Google](https://developers.google.com/custom-search/v1/overview)). Microsoft killed the Bing Search API in mid-2025. Every team that would have reached for an official API is now evaluating exactly the eight providers in this table, which is why pricing moved so much in 2026: Firecrawl cut its Standard plan from $198 to $83/mo on July 29, 2026 ([billing](https://docs.firecrawl.dev/billing)), and Brave killed its free tier in February 2026. ## Why "per 1,000 queries" is a unit trap The single most expensive mistake in this market is comparing headline per-1k numbers across billing models. There are four distinct units hiding behind "per 1k," and the same workload lands at a different number under each.
Per query One call, one unit. Keiro bills this way (1 credit per `/search/fast`, 3 for `/search/content`), and Brave bills per request. Simple to model: multiply queries by the per-1k rate. What you see is close to what you pay.
Per credit, with a multiplier Firecrawl and Tavily bill this way, and the multiplier is where the surprise lives. Firecrawl: 1 credit per basic scrape, +4 for JSON, +4 for Enhanced, so 9 credits with both; a search is 2 credits per 10 results. Tavily: 1 credit for basic search, 2 for advanced, 15 to 250 for a single deep-research request. The per-credit rate is the floor; the per-operation multiplier is the bill.
Per request plus tokens Perplexity Sonar charges $5 to $14 per 1,000 requests by model and context depth, then adds token fees on top, $1/$1 per 1M for Sonar, $3/$15 for Sonar Pro ([pricing](https://docs.perplexity.ai/docs/getting-started/pricing)). The request fee is the floor; the token bill is the variable that grows with answer length.
Per search, volume-tiered Serper and SerpAPI sell prepaid credit/search pools at tiered rates. Serper: $1/1k at 50k credits, $0.30/1k at 12.5M. SerpAPI: $25/1k at the Starter plan, $9.17/1k at Big Data. The tier you land on depends on the commitment, not the workload.
**The right way to compare is to convert every provider to the cost of 1,000 of your actual call type.** If your agent does a search and then reads the top page, the unit is "search-plus-content." Serper's $1/1k search becomes $1 plus a $2/1k extraction service plus an embedding step. Keiro's `/search/content` is 3 credits flat. Firecrawl's search-plus-scrape is 3 credits at basic depth, or 11 with JSON+Enhanced. Until you run that conversion, the headline table above is marketing, not engineering. ## What is the per-operation credit math for each API? This is the table the headline comparisons skip. Every credit-based API converts operations to credits at different rates, and the multiplier is a bigger cost driver than the per-credit price. | API | Operation | Cost in credits | Effective $/1k | | --- | --- | --- | --- | | **Keiro** | `/search/fast` | 1cr | $1.00 ([pricing](/pricing)) | | **Keiro** | `/search/content` (search + markdown + embeddings) | 3cr | $3.00 | | **Keiro** | `/extract` (content from a URL) | 3cr | $3.00 | | **Keiro** | `/answer` (synthesized, cited) | 5cr | $5.00 | | **Keiro** | `/agentic` (deep research) | 20cr | $20.00 | | **Keiro** | `/batch` (async, per query) | 1cr | $1.00 | | **Firecrawl** | search | 2cr / 10 results (rounded up) | $1.66 | | **Firecrawl** | scrape, basic | 1cr / page | $0.83 | | **Firecrawl** | scrape, JSON + Enhanced | 9cr / page | $7.47 | | **Tavily** | basic search | 1cr | $5-8 | | **Tavily** | advanced search | 2cr | $10-16 | | **Tavily** | extract | 1cr / 5 URLs | ~$1-1.60 | | **Tavily** | research (pro model) | 15-250cr | $75-2,000 | | **Exa** | search | $7 / 1k req | $7.00 ([pricing](https://exa.ai/pricing)) | | **Exa** | contents | $1 / 1k pages | $1.00 | | **Exa** | extra results past 10 | $1 / 1k results |, | | **Perplexity** | Sonar request, low/med/high | $5 / $8 / $12 per 1k | + tokens ([pricing](https://docs.perplexity.ai/docs/getting-started/pricing)) | | **Perplexity** | Sonar Pro request | $6 / $10 / $14 per 1k | + $3/$15 tokens | The takeaway from the credit math is that **the same workload can differ 4x across providers purely from operation multipliers.** Firecrawl's headline $0.83/1k credits is the cheapest per-credit rate in the set, but a RAG page scrape with JSON and Enhanced Mode is 9 credits, and 11 with the search call that found the page. Tavily's research endpoint at 15-250 credits per request is a budget wildcard no headline can capture. Keiro's credit costs are 1, 3, 5, and 20, the deep-research endpoint is its most expensive operation, which is worth knowing, but the operation you actually run in a RAG loop (`/search/content` at 3cr) is the second-cheapest content-bearing call in the market. ## Which provider is cheapest per 1,000 searches? Provider by provider, with the numbers from the August 2026 pricing pages. Where a provider wins for a specific workload, the section says so plainly.

Keiro index + content, one call

Price ~$1/1k search at Startup ($50/mo for 50,000 credits, 1,000 req/min). Free tier 500 credits/mo, 30 req/min, no card. Essential $15/mo for 5,000cr, Pro $25/mo for 15,000cr.
Unit per query. 1cr `/search/fast`, 3cr `/search/content` (search + clean markdown + optional embeddings), 3cr `/extract`, 5cr `/answer`, 20cr `/agentic`, 1cr/query `/batch`.
The catch Newer SDK ecosystem than the incumbents. Index-first (keyword plus freshness), not keyword-free semantic "find similar." Not the accuracy leader on every benchmark, FinanceBench puts it at 78%, respectable but not top.
Keiro is the cheapest full-featured option in this set, and the reason is structural, not promotional. **The 3-credit `/search/content` call replaces a search-plus-scrape-plus-embed stack, so the effective cost of a full RAG query is about $3/1k with no second service.** On the public FinanceBench set Keiro scores 78%, honest, not the leader, but this comparison is about cost, and no other provider returns search results, clean page text, and embeddings in a single 3-credit call. The free tier (500 credits/mo, no card) covers about 166 full `/search/content` queries a month, which is enough to port a real query stream before you pay.

Pros

Cons

Serper cheapest raw SERP

Price from $1/1k (Starter, 50k credits for $50), down to $0.30/1k (Ultimate, 12.5M for $3,750). 2,500 free credits one-time.
Unit per search. 1 credit = up to 10 results; 11 to 100 results = 2 credits. Credits valid 6 months, do not roll over.
The catch Raw SERP only. No content, no answers, no AI. You build the extraction layer, and the $0.30/1k volume tier is at 12.5M credits.
Serper wins the raw-SERP category outright. **Sub-100ms, dead-simple REST, prepaid credits, and a volume tier that falls to $0.30/1k, if your workload is "give me Google results and I'll parse them myself," nothing else in this list touches it.** The honest read is that it is a component, not a search-and-reason API. Pair it with an extractor for RAG and you now run two services with two failure modes and two invoices. The TCO including the extractor and glue code is higher than the $1/1k headline suggests, the one-call-vs-two-call section prices this out.

Firecrawl full scrape suite, credit traps

Price $0.83/1k credits at Standard ($83/mo for 100,000 credits, a 58% cut from $198 on July 29, 2026). 1,000 free credits/mo.
Unit per credit. Scrape 1cr/page, crawl 1cr/page, map 1cr/call, search 2cr/10 results (rounded up), interact 2cr/browser-minute.
Add-ons JSON extraction +4cr/page, Enhanced Mode +4cr/page. Scrape with JSON+Enhanced = 9cr/page. Search + JSON+Enhanced scrape = 11cr per result page.
The catch The 1-credit-per-page headline becomes 9 credits with the features you actually need. Search credits round up. Smart Upgrade (June 2026) auto-tiers you up when balance hits zero.
Firecrawl is the cheapest per-credit provider and the most expensive per-operation one, simultaneously. **A basic scrape is 1 credit; turn on JSON extraction and Enhanced Mode and the same page is 9 credits, a 9x multiplier on the headline.** A search is 2 credits per 10 results, rounded up, so 11 results costs 4 credits. The July 2026 price cut makes the credit rate genuinely cheap, but only if your workload stays on basic operations. For extraction-heavy workloads against JS-heavy pages, Firecrawl is the specialist; for search-plus-content RAG, the credit stack adds up faster than the headline suggests.

Tavily agent-native, plan tiers matter

Price $5-8/1k depending on plan. PAYG $0.008/cr ($8/1k). Growth: 100k credits for $500/mo ($5/1k). Startup: 38k for $220 ($5.80/1k). Bootstrap: 15k for $100 ($6.70/1k). Project: 4k for $30 ($7.50/1k).
Unit per credit. Basic search 1cr, advanced 2cr. Extract 1cr/5 URLs (not charged for failures), advanced extract 2cr/5 URLs. Map 1cr/10 pages. Research 15-250cr (pro), 4-110cr (mini).
The catch The $8/1k PAYG rate is the ceiling; the $5/1k Growth rate requires a $500/mo commitment. Research at 15-250cr per request is a budget wildcard.
Tavily is the most plan-dependent price in this set. **Pay-as-you-go at $0.008/credit is $8/1k for a basic search; the Growth plan drops it to $5/1k but only once you commit to $500/mo.** The 1.6x gap between the two is entirely a commitment decision. The per-operation model is fair in places, extract is 1 credit per 5 successful URLs and failures are not charged, but the research endpoint's 15-to-250-credit range on the pro model is a 16x spread on a single call, and a research-heavy agent can blow through a month of credits on one deep run. Credits reset monthly; they do not roll over.

Brave independent index, no free tier

Price $5/1k requests (Search plan). No free tier (removed February 2026). Answers plan $4/1k + $5 per 1M input and $5 per 1M output tokens.
Unit per 1k requests.
Rate limit up to 50 QPS.
The catch The free tier is gone; a $5/mo credit remains but only with public attribution. LLM Context API included, but full-page RAG may still need a separate extractor.
Brave is the only independent Western search index left after Bing's shutdown, and it prices accordingly. **At $5/1k it is 5x Keiro's search rate, and the free tier that existed as late as 2025 is now $5 in monthly credits that require public attribution** ([pricing](https://brave.com/search/api/)). The LLM Context API, structured, token-efficient chunks for agents, is included in the Search plan, which is a real differentiator for agent workloads. But it returns snippets and context, not full page text, so a RAG pipeline that needs to read articles still adds a separate extraction step on top of the $5/1k.

Exa semantic, biggest free tier

Price $7/1k search. Contents $1/1k pages per content type. Extra results past 10: $1/1k each. Answer $5/1k. Deep Search $12/1k, Deep-Reasoning $15/1k. Monitors $15/1k.
Free tier roughly 20,000 free search requests/mo ($20 signup credits plus $10/mo in recurring credits) ([pricing](https://exa.ai/pricing)).
Rate limit 10 QPS search, 100 QPS contents. Enterprise for more.
The catch The a-la-carte menu adds up. A search + contents + answer pipeline is $7 + $1 + $5 per 1k of each. 10 QPS is tight for a busy agent stack.
Exa has the biggest free tier in the set and the clearest a-la-carte pricing, and the two facts pull in opposite directions. **The 20,000 free search requests/mo cover most prototyping and light production; the moment you pay, contents cost $1/1k pages, extra results past the tenth cost $1/1k each, and the menu stacks.** Since March 2026, contents for the first 10 results are bundled into the $7 search, a real improvement, but a search that pulls 30 results for RAG re-ranking effectively bills $27/1k once the extra-result charges land. The real differentiator is keyword-free neural "find similar" search, which none of the index-first APIs replicate, and the honest caveat is that its semantic depth is not what this cost comparison rewards.

Perplexity (Sonar) cited answers, token trap

Price Sonar $5 to $12/1k by context depth plus $1/$1 per 1M tokens. Sonar Pro $6 to $14/1k plus $3/$15 per 1M. Reasoning Pro $6 to $14/1k plus $2/$8. Standalone Search API $5/1k (no tokens).
Unit per request (by depth) plus per token.
The catch The token fee is the hidden cost. The $5/1k headline understates real spend, and there is no permanent free tier.
Perplexity is the strongest option when you want a finished, cited answer in one call and do not mind paying per token for it. **On a short factual query, the $0.005 request fee usually outweighs the token cost; on a long research answer, the token bill can rival it** ([pricing](https://docs.perplexity.ai/docs/getting-started/pricing)). Sonar Pro's $15/1M output rate is the single biggest per-token cost in this set. The standalone Search API at $5/1k with no token cost is the cheaper path if you only need raw results, and the Agent API's `web_search` tool at $0.0025 per invocation ($2.50/1k) is the entry point for multi-tool agents. But for a RAG pipeline that wants clean content to chunk, Perplexity returns an answer, not raw text to rerank, a shape mismatch, not just a price one.

SerpAPI SERP fidelity, most expensive

Price Starter $25/1k (1k for $25/mo). Developer $15/1k (5k for $75). Production $10/1k (15k for $150). Big Data $9.17/1k (30k for $275). 250 free/mo.
Rate limit up to 100 QPS.
The catch Metadata only. You build the extraction pipeline. 10 to 15x a focused search API at most volumes.
SerpAPI is the most established provider here and the most expensive by a wide margin. **You pay for 40+ search engines, a 99.9% SLA, up to 100 QPS, and rich structured SERP, ads, knowledge panels, featured snippets, and the cheapest plan is $25/1k.** The rate falls 2.7x to $9.17/1k at Big Data, but only once you are buying 30k searches/mo, and even that floor is roughly 9x Keiro's search rate. It is metadata only, so the extraction layer for LLM use is your problem. The right pick when you need Google SERP verticals (shopping, maps, images) the AI-native APIs do not surface and the cost is absorbable. ## What does each provider cost at 10k / 100k / 1M queries a month? This is the table that matters for anything past prototyping. Entry rates, applied at volume, no hidden costs yet. | Provider | $/1k (entry) | 10k/mo | 100k/mo | 1M/mo | | --- | --- | --- | --- | --- | | **Keiro** | $1.00 | $10 | $100 | $1,000 | | **Serper** | $1.00 | $10 | $100 | $1,000 (entry) / $300 (volume) | | **Firecrawl** (search) | $1.66 | $17 | $166 | $1,660 | | **Brave** | $5.00 | $50 | $500 | $5,000 | | **Perplexity** | ~$5 + tokens | $50+ | $500+ | $5,000+ | | **Tavily** | $5-8 | $50-80 | $500-800 | $5,000-8,000 | | **Exa** | $7.00 | $70 | $700 | $7,000 | | **SerpAPI** | from $15 | $150 | $1,500 | $9,170-15,000 |
Monthly cost at 1 million queries per month across eight providers
At 1M queries/mo, Keiro is about $1,000 and SerpAPI is $9,170 to $15,000. Same volume, a 9x to 15x spread before hidden costs.
The spread at volume is the number to internalize. **At 1M queries/mo, the cheapest full-featured option is about $1,000 and the most expensive is $15,000, a 15x spread for the same search volume, before a single hidden cost.** At 100k/mo the gap is $100 to $1,500. Most teams choosing between these providers are choosing at 100k to 1M/mo, and that is exactly the range where the headline differences compound into budget differences. ## Where do the volume-tier cliffs hide? The single biggest pricing lever in this market is volume tiering, and it is where misbudgeting happens. Incumbents discount only after a large monthly commitment; the effective rate drops the more you commit, and the drop is gated behind a spend threshold.
SerpAPI Entry $25/1k (Starter) falls to $9.17/1k at Big Data (30k for $275/mo). A 2.7x drop, only once you are buying 30k/mo, a $275/mo commitment just to unlock the rate.
Serper Falls from $1/1k to $0.30/1k once you buy 12.5M credits. A 3.3x drop at a volume most teams never reach, gated by a $3,750 purchase.
Tavily Drops from $8/1k on PAYG to $5/1k on Growth (100k credits for $500/mo). A 1.6x drop that requires a $500/mo commitment.
Keiro $1/1k flat across tiers. No cliff, no commitment. The entry rate is the scale rate, and the per-credit rate at Startup ($50 for 50,000cr) is the same $1/1k.
The pattern is consistent: the discount is real, and it is priced behind a commitment you may not want. **A team at 30k searches/mo does not get SerpAPI's $9.17/1k, it pays $10-15/1k until it buys the Big Data plan.** Model your real volume against the tier table before you budget against the best-case rate. At volume, Keiro's flat $1/1k means the rate you test on is the rate you scale on, which is why the cost-at-volume table above shows it sitting at the floor at every row. ## What are the six hidden costs on top of the headline? The advertised per-1k price is the base layer. Six costs stack on top, and they are the reason two APIs with the same headline can differ 3x in production.
1. Separate extraction Serper, Brave, and SerpAPI return metadata only. If your agent needs full page text for RAG, you build or buy a separate extraction step, a second API, a second failure mode, a second invoice. Exa charges $1/1k pages for contents on top of $7/1k search (first 10 results bundled since March 2026). Keiro bundles search and clean text in `/search/content` for 3 credits. Tavily includes content inline and has a dedicated `/extract` (1 credit per 5 successful URLs).
2. Credit multipliers The one that catches teams. Firecrawl's "1 credit per page" becomes 9 credits with JSON (+4) and Enhanced (+4), and 11 with the search that found the page. Tavily's research endpoint costs 15 to 250 credits per request on the pro model. If you compared Firecrawl's $0.83/1k credits to Keiro's $1/1k and picked Firecrawl, your real cost may be higher because your workload ran the 9-credit operations, not the 1-credit ones.
3. Token costs Perplexity Sonar charges $5 to $14 per 1k requests by model and depth, plus $1/$1 per 1M tokens (Sonar Pro: $3/$15). A 500-input, 200-output query is about $0.005 in tokens on top of the $0.005 request fee; a 2,000-output-token research answer adds about $0.002. At scale the token bill can rival the request bill, and the "$5/1k" headline understates spend by the entire token variable.
4. Rate-limit upgrades Every provider throttles at a per-minute or per-second limit tied to your tier. Exa caps search at 10 QPS with no published tier system below Enterprise. Keiro's free tier is 30 req/min, Startup 1,000/min. Firecrawl limits by concurrent browsers. Exceed the limit and requests queue or fail; the fix is a plan upgrade, and the upgrade cost is real. Budget for the tier your peak, not your average, requires.
5. Plan minimums and commitments SerpAPI's cheapest plan is $25/mo whether you run 100 searches or 1,000. Tavily's $5/1k rate requires a $500/mo Growth commitment; below that you pay $5.80-8/1k. Serper's volume discounts are prepaid purchases ($3,750 for the $0.30/1k rate), not subscriptions. The commitment is part of the price.
6. Free-tier strings Brave removed its free tier in February 2026; a $5/mo credit remains, but only with public attribution. Exa's free tier is search only, contents, answers, and deep search cost credits. Serper's 2,500 free credits are one-time, not recurring. Firecrawl's 1,000 free credits go fast at 9 credits per JSON+Enhanced page. Read the free-tier terms before you architect around them.
**The credit-vs-query unit trap is the root of most of these surprises.** "Per 1k" is not comparable across a per-query API, a per-credit API, and a per-request-plus-token API until you convert to your real call shape. A team that compared Firecrawl's $0.83/1k credits to Keiro's $1/1k queries and picked Firecrawl found it cost more, because its workload was search calls (2 credits each, rounded up) and JSON+Enhanced scrapes (9 credits each), not basic page scrapes. ## What does one call cost versus two calls for RAG? The real cost of a RAG query is search plus the page text the agent reads plus embeddings, if you want the chunks ready for a vector store. Some providers sell that as one call; most sell it as two or three. This is the comparison that determines your bill.
Stacked bars of true RAG cost per 1,000 queries: one call versus two-call stacks
True RAG cost per 1k: Keiro is a single $3 call. Every other provider stacks a search, an extraction step, and sometimes tokens on top.
For a workload that searches, reads the top page, and embeds it, the all-in per-1k costs are roughly: Keiro $3.00 (one call, 3cr), Firecrawl $2.79 (search 2cr + scrape 1cr + embeddings, before JSON/Enhanced multipliers), Serper $3.30 (search $1 + a $2/1k extractor + $0.30/1k embeddings), Tavily $5.30 on Growth (search with content + embeddings), Perplexity $5.70 (request plus tokens), Brave $7.30 (search + extractor + embeddings), Exa $7.30 (search with bundled first-10 contents + embeddings), SerpAPI $17.30 (search + extractor + embeddings). **Keiro at $3/1k and Serper-plus-extractor at $3.30/1k land at the same number, the difference is that Keiro is one call with one invoice, and the Serper stack is two services you wire together.** Firecrawl undercuts both on paper at basic depth, then jumps to roughly $9.10/1k the moment JSON extraction and Enhanced Mode are on. The two-call stacks also hide an engineering cost: merging two response shapes, handling two rate limits, and keeping two vendors' schemas in sync is real work that does not appear on any invoice but shows up in your roadmap. ## What is the true cost of ownership for a RAG pipeline? The API bill is not the total bill. The per-1k price covers the API call; it does not cover the infrastructure you build around it.
Separate extraction Serper, Brave, and SerpAPI return metadata. For RAG you need clean page text, so a second API (Jina Reader at roughly $0.05/1M tokens, or a self-hosted scraper) plus glue code to orchestrate two services, handle two failure modes, and merge two response shapes. Add $1 to $5 per 1k queries, plus engineering time.
Embedding service RAG needs embeddings. Keiro bundles them in `/search/content` (3 credits, optional via `embeddings.enabled`). Most providers do not, so you run a separate embedding model on the clean text, OpenAI, Cohere, or a local model. Add $0.10 to $0.50 per 1k, plus the infrastructure to run it.
Caching layer Firecrawl's cached results still cost 1 credit per page; caching is speed, not savings. To avoid paying twice for the same query you build your own cache in front of the API. Infrastructure on top of the API cost.
Retry and fallback APIs fail, rate limits hit, timeouts happen. Production agents need retry logic with backoff and a fallback provider. Engineering cost, not API cost, but real and ongoing.
The TCO gap is where the one-call providers pull away. **A SERP-only API at $1/1k plus a $2/1k extractor plus a $0.30/1k embedding service is $3.30/1k all-in, plus the engineering to wire three services.** Keiro at $3/1k includes search, clean text, and embeddings in one call, no second service. The gap is not just the API price; it is the infrastructure and engineering the per-1k headline does not cover. ## Where does Keiro win, and where does it lose? A pricing comparison that only lists wins is a pamphlet. Here is where Keiro wins, where it loses, and the specific workloads where each competitor is the right answer.
Cheapest full-featured (wins) At ~$1/1k search and ~$3/1k all-in via `/search/content`, Keiro is the cheapest provider that returns search results, clean content, and embeddings in one call. At 1M queries/mo that is about $1,000; the next cheapest content-bearing option (Tavily on Growth) is $5,000 at the same volume.
Raw SERP at extreme volume (loses) Serper at $0.30/1k at 12.5M credits ([pricing](https://serper.dev)) is 3.3x cheaper than Keiro's $1/1k for raw SERP-only workloads. If you just need Google results and parse them yourself, Serper is cheaper, at 12.5M queries/mo that is $3,750 versus $12,500.
Biggest free tier (loses) Exa's roughly 20,000 free search requests/mo ([pricing](https://exa.ai/pricing)) is 40x Keiro's 500 credits/mo for raw search volume. For prototyping discovery-only workloads, Exa's free tier covers far more queries, with the caveat that contents, answers, and deep search all cost credits.
Independent privacy index (loses) Brave runs its own 30B+ page index, SOC 2 Type II, no tracking ([pricing](https://brave.com/search/api/)). At $5/1k it is 5x Keiro's rate, but if your requirement is a privacy-guaranteed independent index, price is not the decision axis.
Full crawl suite (loses) Firecrawl's scrape, crawl, map, search, interact, and extract in one API is broader than Keiro's indexed-search-plus-extract surface. If your workload is crawling a known site at depth with JS rendering, Firecrawl is the specialist.
The honest summary: **Keiro wins the "cheapest full-featured option" title because it is the only provider where search, content, and embeddings are one $3 call. Every category win it concedes, raw SERP, free-tier size, independent index, crawl suite, is a workload where the unit economics genuinely favor someone else.** That is the nature of a real comparison, and it is why the per-1k headline alone will not tell you which API to use. ## How do you choose for your budget?
Sub-$100/mo (prototyping) Stay on free tiers first. Exa's 20,000 free search requests cover most discovery prototyping. Tavily's 1,000 credits and Keiro's 500 credits are enough to port a real query stream; Keiro's 500 credits cover about 166 full `/search/content` queries. Serper's 2,500 one-time credits last a while for raw SERP. When you go paid, Keiro at ~$1/1k means 100k queries is about $100, the cheapest full-featured floor.
$100 to $1,000/mo (growing) Keiro is the cheapest full-featured option: 100k queries is about $100, bundled extraction and embeddings. Serper at $1/1k is the same price only if you need raw SERP and will build extraction. Tavily's Growth plan at $500/mo for $5/1k is competitive if you are committed to LangChain. Firecrawl is competitive if your workload is mostly basic scrapes, watch the multipliers the moment JSON or Enhanced Mode turn on.
$1,000+/mo (scale) Volume tiers start to matter. SerpAPI drops to $9.17/1k at Big Data, Serper to $0.30/1k at 12.5M credits, Tavily caps at $5/1k on Growth. Keiro stays ~$1/1k flat. At a million queries/mo, Keiro is about $1,000 and the next cheapest full-featured option is several times that. Model your real volume against each provider's tier table, not the headline.
Free-tier-first Exa ~20k (search only, most generous), Tavily 1k, Firecrawl 1,000 credits, Keiro 500, SerpAPI 250, Serper 2,500 one-time. Brave has no free tier anymore. Do not architect around a free tier without reading its strings: Exa's is search-only, Brave's credit needs attribution, Serper's is one-time, and Firecrawl's 1,000 credits go fast at 9 credits per JSON+Enhanced page.
The cheapest one-call API that returns clean content with embeddings. 500 free credits/mo, no card, port your real query stream before you pay. Start free · 500 credits/mo See pricing
## Further reading For the full provider-by-provider tables, per-operation credit math, and independent breakdown behind this comparison, see [Web Search API Comparison 2026: 12 Providers](https://www.agenticsearch.cloud/posts/search-api-comparison/) and [Best AI Search APIs in 2026: Cheapest for RAG](https://www.agenticsearch.cloud/posts/best-ai-search-api/). If you are building the data pipeline that feeds these APIs, [Scraping at scale: queues, caching, and not getting banned](https://www.webscraping.space/blog/scraping-at-scale) covers the infrastructure the per-1k headline does not. More from this series: - [Top Web Search APIs (Aug 2026)](/blogs/comparisons/top-web-search-apis-aug-2026) - [Best Free Web Search API (Aug 2026)](/blogs/guide/best-free-web-search-api-aug-2026) - [Best Web Search API for AI Agents (Aug 2026)](/blogs/guide/best-ai-search-apis-for-agents-aug-2026) - [Cheapest Web Scraping APIs (Aug 2026)](/blogs/guide/cheapest-web-scraping-apis-aug-2026) - [Keiro pricing](/pricing), the exact credit math behind the per-1k figures ## FAQ ### What is the cheapest web search API in August 2026? For full-featured search with bundled content extraction and embeddings, Keiro at ~$1/1k search (about $3/1k all-in for a RAG query via `/search/content`, 3 credits). For raw Google SERP only, Serper from $1/1k, down to $0.30/1k at 12.5M credits, you build extraction. For extraction-heavy scraping, Firecrawl at $0.83/1k credits, but only if you stay on basic scrapes (1 credit each); JSON+Enhanced is 9 credits per page. "Cheapest" is the wrong question without "cheapest for what operation." ### What does Keiro cost per 1,000 queries? ~$1/1k for search on the Startup tier ($50/mo for 50,000 credits, 1,000 req/min). The `/search/content` endpoint (search plus clean markdown plus optional embeddings) is 3 credits, about $3/1k all-in. `/extract` is 3cr, `/answer` is 5cr, `/agentic` deep research is 20cr, and `/batch` is 1cr/query. The Free tier is 500 credits/mo with a 30 req/min limit and no card. ### What is the cheapest raw SERP API? Serper, from $1/1k at the Starter tier (50k credits for $50), falling to $0.30/1k at the Ultimate tier (12.5M credits for $3,750). It returns raw Google SERP in under 100ms with no content, no answers, and no AI. SerpAPI, the other established SERP option, is from $15/1k at Developer and $25/1k at its cheapest plan. ### What are the per-operation credit costs for Firecrawl? Basic scrape: 1 credit per page. JSON extraction: +4 credits. Enhanced Mode: +4 credits. Scrape with both: 9 credits per page. Search: 2 credits per 10 results, rounded up (11 results = 4 credits). Crawl and map: 1 credit per page or call. Interact: 2 credits per browser minute. At the Standard plan's $0.83/1k credit rate, a search is about $1.66/1k and a JSON+Enhanced scrape is about $7.47/1k pages. ### What are the per-operation credit costs for Tavily? Basic search: 1 credit. Advanced search: 2 credits. Extract (basic): 1 credit per 5 successful URL extractions, not charged for failures. Extract (advanced): 2 credits per 5 URLs. Map: 1 credit per 10 pages. Research (pro model): 15 to 250 credits per request. Research (mini model): 4 to 110 credits. Plans range from Researcher (free, 1,000 credits) to Growth ($500/mo, 100,000 credits), with PAYG at $0.008/credit. ### What hidden costs should I budget for beyond the headline price? Per-operation credit multipliers (Firecrawl's 9cr/page with JSON+Enhanced, Tavily's 15-250cr research), content extraction as a separate paid call (Exa contents, or building your own for SERP-only APIs), token fees on top of per-request pricing (Perplexity, worse on Sonar Pro at $3/$15 per 1M), rate-limit upgrades for peak traffic, plan minimums and commitments (SerpAPI's $25/mo floor, Tavily's $500/mo Growth commitment), free-tier strings (Exa search-only, Brave attribution-required, Serper one-time), and the credit-vs-query unit trap. Add these to the headline before you commit. ### What is the true cost of ownership for a RAG pipeline? The API bill plus the infrastructure around it. SERP-only APIs (Serper, Brave, SerpAPI) need a separate extractor ($1-5/1k) plus glue code. RAG needs an embedding service ($0.10-0.50/1k) unless the API bundles embeddings, Keiro does, in `/search/content`. Add a caching layer and retry/fallback logic for production. A SERP-only API at $1/1k plus a $2/1k extractor plus $0.30/1k embeddings is $3.30/1k all-in across three services. Keiro at $3/1k is one call, no second service. ### Which API has the biggest free tier? Exa, with roughly 20,000 free search requests/mo ($20 signup credits plus $10/mo recurring). It is search only, contents, answers, and deep search cost credits. Tavily gives 1,000 credits/mo, Firecrawl 1,000 credits/mo, Keiro 500 credits/mo (no card), SerpAPI 250 searches/mo, and Serper 2,500 credits one-time. Brave's free tier was removed in February 2026. ### Why is SerpAPI so expensive? You are paying for 40+ search engines, a 99.9% SLA, up to 100 QPS, and rich structured SERP (ads, knowledge panels, featured snippets), not for cheap queries. It is metadata only, so the extraction layer is yours to build. The rate drops to $9.17/1k at Big Data (30k for $275/mo), but at most volumes it is 10 to 15x a focused search API, and the $25/1k Starter plan is the cheapest entry. ### How do I compare search API pricing accurately? Convert every provider to the cost of 1,000 of your actual call type: basic search, search-plus-content, search-plus-content-plus-embeddings, answer, or deep research. Add per-operation credit multipliers (Firecrawl's 9cr/page, Tavily's 15-250cr research). Add hidden costs (separate extraction, tokens, rate-limit upgrades, embeddings). Apply your real volume to each provider's tier table (SerpAPI drops 2.7x at Big Data, Serper 3.3x at 12.5M, Tavily 1.6x on Growth, Keiro flat at ~$1/1k). Model a full month. The headline per-1k rate alone will mislead you on Firecrawl, Tavily, Perplexity, and SerpAPI every time.