Keano

💰 Tired of SERP APIs priced like enterprise SaaS? SerpBase: 0.30/1,000searches.0.30/1,000searches.3 gets you 10,000. No monthly fees. Credits never expire. Auto top-up kicks in when your balance runs low — no API downtime at 3 AM. Real-time Google Search API. Clean JSON. 200+ countries. 1.4s avg latency. If you're building rank trackers, AI agents, or data pipelines, the math is simple. 100 free searches, no card required.

@houzhongji89090 gpt生成的，推特上用真人头像好像回复率要高点😂

@serpbase 你这个头像

推荐一个给 Agent 用的资源层：AIsa.one 只要给小龙虾、Hermes 这类 Agent 丢一段提示词，就能快速接上各种大模型、数据 API 和现成 Skill。 比如 DataForSEO、Twitter/X、Apollo，Agent 可以直接调用。 适合需要做自动化工作流、增长 Agent、SEO Agent、营销 Agent 的人。

@houzhongji89090 好家伙，居然是熟人🥲

@serpbase 你这个思路很有想法

@B2the7 The gap between Google rankings and AI citations is becoming the real SEO headache. You can rank and still not get cited, or get cited from pages that don’t look important in classic SERPs.

44.2% of AI citations come from the first 30% of your content. Bury your answer and AI skips you entirely. Learn how ChatGPT, Gemini, and Perplexity each select sources differently b2the7.com/news-blog/goog… #GEO #ContentStrategy

@KendraLeeKLA The tricky part is that “AI visibility” sounds abstract until you tie it back to actual search surfaces: Google results, News, brand pages, reviews, citations. That’s where structured SERP data helps.

AI search changed visibility. Buyers now ask ChatGPT, Gemini, and Perplexity before visiting websites. If AI doesn’t trust your brand enough to recommend it, you lose visibility before the click. See how your brand performs: hubs.li/Q04h-QrZ0 #AEO #AISEO

@justbyte_ This is exactly why cheaper Maps data layers need to exist. Most people don’t need the whole spaceship, they just need clean place results, coordinates, phones, websites, and hours.

“Google Maps API is expensive, let’s launch satellites.”

@manol_ai If you keep testing sources, I’d separate “discovery” from “enrichment.” Google Maps for discovery, then email/owner enrichment as a second step. SERP/Maps APIs are strongest upstream.

Every lead source I've tested this month, ranked by reply rate: DataForSEO (peptides): 0.52% reply, 5 opps Google Maps + OpenWebNinja (med spas): 0.80% reply, 6 opps Apollo (B2B agencies): 0.11-1.45% (varies by niche) Apify scraped (various): 0.10-0.41% Maps + OpenWebNinja wins for local. DataForSEO best for niche B2B. #buildinpublic #b2bsales #leadgeneration #coldemail

Good news: our site is back in Bing’s index 🎉 Also discovered Bing has a Rewards program while checking everything, and somehow ended up redeeming a gift card. Not the recovery workflow I expected, but I’ll take it. #SEO #Bing #BingSEO #TechnicalSEO

SEO people, has anyone dealt with Bing just nuking a site from the index? Bing Webmaster Tools shows no manual action, no crawl issues, nothing obvious. Our best guess is a malicious report because this happened after we launched very low pricing in a competitive space, but we can’t prove it. Any tips for getting Bing to re-crawl / re-evaluate faster? #SEO #TechnicalSEO #Bing #BingSEO #SearchEngineOptimization

@oleksoleksoleks yeah that’s the part people skip. “breakeven before error rates” is doing a lot of work here

@serpbase They're not cheaper. I worked on an AI project to replace a bunch of our lower-skilled white collar workers, and even then it came to about breakeven with fully loaded cost -- if we neglected the error rates

“AI is cheaper than employees” is one of those sentences that sounds true until you add: inference cost retries hallucination review workflow integration security human oversight the guy whose entire job is now checking what the AI did suddenly the spreadsheet gets a little less magical

@ProSamik Furthermore, Serpbase points never expire, there's no minimum monthly spend, and payments are automatically deducted based on usage. If you'd like to try it out, I'd be happy to offer you a trial opportunity.

@serpbase how is it different than dataforseo?

So today I have been trying to make my SEO process little Agentic using Connectors Here if you see- - Google Analytics for tracking page visits - Search Console for Google ranking - OpenSEO for Competitor Analysis and Backlink analysis - Web Search and SERP API (from openseo) for research on keyword - Subclip Blogs (payload CMS) to write content there I will manually try to experiment few days with this setup and see how it turns out.

@ProSamik SerpBase focuses on Google search, boasts extremely fast response times, returning results in an average of 1.4 seconds, with a success rate of up to 99.9%, and costs only $3 for 10,000 requests.

@beyondGodShit MCPでSERP APIを挟むなら、返却JSONの形が安定しているかがかなり大事ですね。エージェント側のパース処理が軽くなります。

MCPについて少しづつわかってきた。これが「接続」ってことか OpenCodeアプリのkimi2.6ではWEBサーチができないが、SERP API経由のMCPサーバーを構築させることで、WEBサーチを可能とした。 MCPとはトンネルってことなんだな

@ankuuushdev If you want to compare options later, SerpBase is worth testing for the Google Search layer: POST JSON, API key auth, and structured result arrays without maintaining a scraper.

Week 3 Day 3: Built a research system today. • Researcher agent → uses Serper (Google Search API) to gather company info • Analyst agent → takes research output and generates the report Research → analysis → report. #AI #AIAgents #BuildInPublic

@JulianGoldieSEO I have some websites that specialize in SEO testing, and I've achieved quite good results using this model.

CHINA JUST DROPPED A FREE SEO MODEL THAT SHOULD WORRY EVERYONE And the wild part is most people still have no idea it exists. What Just Happened: → Alibaba quietly released Qwen 3.7 Max Preview and Qwen 3.7 Plus Preview → It appeared on LM Arena without a massive announcement → Qwen is now ranked as the number 6 AI lab for text and number 5 for vision → This came only around 1 month after the previous Qwen release Why SEO People Should Care: ✓ It has a thinking mode that plans before writing ✓ It can generate longtail keyword ideas based on beginner search intent ✓ It can write human-style blog posts with hooks, examples, structure, and CTAs ✓ It can review screenshots for SEO, layout, and conversion problems ✓ It can generate schema markup, meta tags, and small SEO code fixes The Real SEO Play: → Use Qwen for keyword research → Use thinking mode for article planning → Use multimodal uploads for competitor page audits → Use coding prompts for schema and technical SEO Free model. Top 6 ranking. Thinking mode. Vision support. SEO workflows built in. The lesson is simple: Stop only testing AI tools for “cool outputs.” Start testing them for traffic, rankings, and conversions.

starbucks killed its ai inventory tool after 9 months because it kept miscounting basic items which is basically the funniest version of enterprise ai: save 10 minutes doing inventory, spend 45 minutes verifying whether the machine hallucinated oat milk finance.yahoo.com/sectors/techno…

@JohnSM_99 Absolutely right!

@serpbase My whole issue is the unreliability of the eco system. I just don’t trust them to NOT upend my workflow anymore with some major change. I can’t hope just hope and pray they ‘reset usage’ out of the goodness of their hearts each time they f something up.

gemini 3.5 flash is such a google product lol it’s fast enough to make everyone go “wait are they back?” then expensive enough to remind you yes, this is still google then weirdly inconsistent enough that nobody actually trusts it until some guy on reddit burns $40 running his own benchmark spreadsheet at 2am the future is here and it requires a pricing calculator, three eval suites, and emotional damage #gemini

@massy_jimmy @virtualfilmer for example, give it a realtime user monitoring task and it can move fast while still picking the wrong architecture with total confidence

@virtualfilmer @serpbase Like what’s an example that it gets wrong?