Keano @serpbase

Building SerpBase — low-cost, fast Google data APIs for developers. Search, Maps, News, Images, Videos. Non-expiring credits, auto top-up, stable latency. serpbase.dev Joined June 2022

Tweets

201
Followers

77
Following

175
Likes

349

Keano @serpbase

3 days ago

Significant improvement, average latency reduced from 1.42 seconds to 0.76 seconds！

0 0 2 24 0

View Details

@tinomadzungu We offer not only map search services but also search engine results pages (SERPs). The SERP API costs approximately 80 times more than our service. We highly recommend you try it out at serpbase.dev

0 0 1 7 0

View Details

Keano @serpbase

5 days ago

@HarveyAix6 If no-subscription is part of the goal, SerpBase’s prepaid SERP API model may be a good fit for the live data layer.

0 0 1 12 0

View Details

Keano @serpbase

5 days ago

@inqodo @elvissun SerpBase could fit the tracking layer if you need Search/Maps JSON with prepaid credits instead of adding another subscription.

0 0 0 10 0

View Details

Keano @serpbase

5 days ago

@matthewsynteria DataForSEO is a solid pick. If you want a lighter prepaid option for Search/Maps JSON, SerpBase could be worth comparing too

0 0 1 35 0

View Details

Keano @serpbase

7 days ago

@cwolferesearch The tool-use layer is where a lot of agent quality comes from. For search specifically, raw browsing and structured SERP APIs solve different problems: browsing is better for deep reading, while SERP data is better for fast grounding and source discovery.

0 0 0 7 0

View Details

Keano @serpbase

7 days ago

@unikoukokun The interchangeable tool interface is the important part. Once search, fetch, code execution, and internal APIs follow a clear contract, teams can swap providers without rewriting the whole agent workflow.

0 0 0 13 0

View Details

Keano @serpbase

7 days ago

@ram_sutraye The “turn off tools you don’t need” point is underrated. Web search is powerful, but it should be called intentionally. For agents, I like separating “need fresh data” from “can answer from local context” before spending tool calls.

0 0 0 1 0

View Details

Keano @serpbase

7 days ago

@oneeyekeh Nice workflow. The research step is probably the part I’d watch most closely: if the web search layer returns noisy or inconsistent results, the whole video pipeline inherits that noise. Structured search results make this kind of automation much easier to debug.

0 0 1 28 0

View Details

Keano @serpbase

7 days ago

@TheTuringPost For agent workflows, I think web search needs to become more like infrastructure: predictable cost, structured output, easy retries. Otherwise the agent spends too much of its budget just figuring out what the web says.

0 0 0 2 0

View Details

Chris Long @chris_nectiv

a week ago

Big SEO News: Google has officially posted an "LLMs.txt" page to the Chrome Developers site. "Without this file, agents may spend more time crawling the site to understand its high-level structure and primary content."

43 110 827 68K 1K

View Details

Keano @serpbase

a week ago

@ClimStefan We develop the interface for the search engine results page (SERP) you might need， please contact me. serpbase.dev

0 0 4 18 0

View Details

Keano @serpbase

a week ago

@allymac81 The economics are interesting. At scale I’d separate keyword lookups from live SERP checks, because they solve different problems and have very different cost profiles.

0 0 4 19 0

View Details

Keano @serpbase

a week ago

@metx_mike @seonatia The nice thing about API-first SEO data is that you can swap the analysis layer. Claude, GPT, local model, doesn’t matter as much if the underlying SERP/keyword data is clean.

0 0 1 34 0

View Details

Keano @serpbase

a week ago

@ClimStefan @csaba_kissi Makes sense to build this around an API. I’d still keep the SERP provider swappable, because rank tracking cost can grow quickly once users add markets and devices.

1 0 1 20 0

View Details

Keano @serpbase

a week ago

claude opus 4.8 is funny because anthropic is like “sharper judgment, more honesty, longer independent work” and the entire subreddit is just sitting there clutching opus 4.6 like a family heirloom #Claude #ClaudeAI #AI #Anthropic

0 0 4 86 0

View Details

Keano @serpbase

a week ago

AI companies keep saying “the future is coming” like that’s supposed to be comforting meanwhile the people graduating into that future are booing AI speakers on stage because the job market already feels like a group project where one guy showed up with a chainsaw #AI #GenZ #FutureOfWork #Tech

1 0 4 105 0

View Details

Keano @serpbase

a week ago

@ProRankTracker Interesting space. Feels like GEO tools will need both classic rank tracking and cheaper SERP collection, because AI visibility checks can multiply query volume fast.

0 0 2 26 0

View Details

Keano @serpbase

a week ago

@laupixagent If this runs repeatedly across SaaS domains, SERP cost becomes part of the product margin pretty quickly. Worth keeping that layer cheap and swappable.