←
Wang shuai (Dylan)183 posts
Wang shuai (Dylan) avatar

Wang shuai (Dylan)

@dylan_wangs

Member of Technical Staff @KeenableAI | Previous Research Fellow at @Ielabgroup | @UWAEMS alumni RAG, IR, NLP, Model Efficiency BlueSky: dylanshuaiwang

◷ Joined February 2018▧ 19 Photos and videos
Haider. avatar
⟳ Wang shuai (Dylan) retweeted
Haider.@haider1·Nov 12

very impressive that Keenable 1.0 is now #1 on Artificial Analysis and NEEDLE search benchmarks just two months after launch instead of relying on third-party APIs, Keenable owns its search stack and learns from AI agent queries through MCP -- and also adds 100m knowledge cards and dynamic snippets that cut token usage by 50% on research tasks

Keenable 1.0 is live Two months of learning from agent queries made Keenable the best web search API for agents: #1 on Artificial Analysis and NEEDLE. 1.0 adds 100M knowledge cards and dynamic snippets that cut tokens by 50% on research tasks. keenable.ai/blog/keenable-1.…

Post attachment
marcus avatar
⟳ Wang shuai (Dylan) retweeted
marcus@marcusyul·Nov 12

KEENABLE IS NOW THE #1 WEB SEARCH API FOR AI AGENTS. Launched August 25. Two months later: → #1 on Artificial Analysis → #1 on NEEDLE Ahead of providers with far bigger budgets. The reason is simple. They own the whole stack, so every query agents sent since launch made it smarter. That's a team learning faster than everyone else. Full comparison in the table below 👇🏼

Post attachment

Keenable 1.0 is live Two months of learning from agent queries made Keenable the best web search API for agents: #1 on Artificial Analysis and NEEDLE. 1.0 adds 100M knowledge cards and dynamic snippets that cut tokens by 50% on research tasks. keenable.ai/blog/keenable-1.…

Post attachment
Andrey Styskin avatar
⟳ Wang shuai (Dylan) retweeted
Andrey Styskin@styskin·Nov 12

Keenable 1.0 is out. Agents want the original document and pay for every token. 1.0 returns only the part they need, cutting tokens 50% on research tasks, and adds 100M knowledge cards. It’s the new champion in web search for agents, and I’m proud of what we’ve built. keenable.ai/blog/keenable-1.…

Post attachment
elvis avatar
⟳ Wang shuai (Dylan) retweeted
elvis@omarsar0·Nov 12

Keenable is now the #1 web search API for AI agents, ranking first on the Artificial Analysis Search Index. Nice progress from @KeenableAI since August. They built their own search stack, use agent queries to improve it, and move faster than anyone. I think that feedback matters as agents rely more on search.

Keenable 1.0 is live Two months of learning from agent queries made Keenable the best web search API for agents: #1 on Artificial Analysis and NEEDLE. 1.0 adds 100M knowledge cards and dynamic snippets that cut tokens by 50% on research tasks. keenable.ai/blog/keenable-1.…

Post attachment
Mehul Mohan avatar
⟳ Wang shuai (Dylan) retweeted
Mehul Mohan@mehulmpt·Nov 12

Keenable is #1 on Artificial Analysis Search Index, NEEDLE benchmark, and I tried it with my own history (graphic below). What if your agent can "select * from web" in real time and the results are the best possible results you'll see? Keenable 1.0 just launched and includes that, among many other things. I asked it to get my whole history of this year, and the result blew me away

Post attachment

Keenable 1.0 is live Two months of learning from agent queries made Keenable the best web search API for agents: #1 on Artificial Analysis and NEEDLE. 1.0 adds 100M knowledge cards and dynamic snippets that cut tokens by 50% on research tasks. keenable.ai/blog/keenable-1.…

Post attachment
Keenable AI avatar
⟳ Wang shuai (Dylan) retweeted
Keenable AI@KeenableAI·Nov 12

Keenable 1.0 is live Two months of learning from agent queries made Keenable the best web search API for agents: #1 on Artificial Analysis and NEEDLE. 1.0 adds 100M knowledge cards and dynamic snippets that cut tokens by 50% on research tasks. keenable.ai/blog/keenable-1.…

Post attachment
Keenable AI avatar
⟳ Wang shuai (Dylan) retweeted
Keenable AI@KeenableAI·Oct 27

Keenable web search infrastructure is now available through @zeroclick, the platform for selling to AI agents. Ask your agent to build a table of founders, funding rounds, and lead investors for 100 companies. This integration lets it buy $2 of credit, make 500 requests and pay from its own balance without opening a Keenable account or leaving the chat interface. Start at zeroclick.keenable.ai

We've partnered with @KeenableAI to bring its web index to AI agents through a ZeroClick Agent Storefront. Your agent can research a company, filter results by site or date, and read the pages it finds, paying as it goes from a prepaid balance. See how it works: zeroclick.ai/blog/keenable-i…

ZeroClick avatar
⟳ Wang shuai (Dylan) retweeted
ZeroClick@zeroclick·Nov 12

Your agent can research a company with @KeenableAI, filter results by site or date and read the pages it finds as clean text. Our explainer shows how it buys that web access through ZeroClick. Have your agent explore zeroclick.keenable.ai

2:20
Shengyao Zhuang avatar
⟳ Wang shuai (Dylan) retweeted
Shengyao Zhuang@ShengyaoZhuang·Nov 11

Not just startups, but university labs too! If you need search API credits for your RL training experiments, reach out to us!

Building an AI startup? Get $5,000 in Keenable credits for 6 months, plus onboarding and help promoting your launch. The fastest web search for agents: results in under 250 ms at p95 in US East. The credits work across every Keenable product. For startups backed by @Accel and @conviction, teams building something cool, and friends. Apply: keenable.ai/blog/ai-startups…

Wang shuai (Dylan) avatar
Wang shuai (Dylan)@dylan_wangs·Nov 11

So glad to be part of Keenable! Seeing the pace, passion, and dedication of the team every day makes me confident that we're destined to build something truly exceptional and deliver the best possible service to our customers; excited for what's ahead!

Keenable Search 1.0 takes 1st (pro, 83) and 2nd (realtime, 82) on the Artificial Analysis Search Index: ahead of Exa (auto, dynamic highlights) and Perplexity (medium) at 80 We benchmarked both modes of @KeenableAI's Keenable Search 1.0, pro and realtime, in our Stirrup agent harness with Keenable Search as its web_search tool to complete 1700 search-focused tasks. At $4 per 1k searches, Keenable costs about $0.040 per task in both modes, over a third below the board median ($0.063). It is also one of the fastest products we measure: 15.0s (realtime) and 15.5s (pro) per task, 2nd and 3rd on the board. Key benchmarking results for Keenable Search 1.0: ➤ Artificial Analysis Search Index: Keenable (pro) scores 83 and Keenable (realtime) 82, up from 67 for both modes when we first benchmarked Keenable in August, and a ~50-point lift over the same model with no search (33). Both lead Exa (auto, dynamic highlights) and Perplexity (medium) at 80 ➤ DeepSearchQA is Keenable's strongest benchmark: both modes score 88 F1, 1st and 2nd of 26, ahead of Exa (auto, dynamic highlights) at 86. Pro also ranks 1st on BrowseComp (87%, level with Perplexity (medium)) and on AA-Omniscience (73%, level with Exa (auto, dynamic highlights) and Firecrawl) ➤ Fast search: realtime takes 2.3s of search time per task (2nd of 26) and pro 3.0s (3rd), against a board median of 12.5s. Each search returns in about 330ms (realtime) and 390ms (pro) on average Key details: ➤ Setup: GPT-5.6 Luna (medium) in Stirrup, Keenable Search API with mode=realtime and mode=pro ➤ Pricing: $4 per 1k requests in both modes (public pay-as-you-go price) ➤ Cards: We disabled Keenable's search cards for testing (disable_cards=true)

Post attachment
Keenable AI avatar
⟳ Wang shuai (Dylan) retweeted
Keenable AI@KeenableAI·Nov 10

Keenable takes #1 on the @ArtificialAnlys Search Index. Less than three months after opening to the public, and three weeks after taking #1 on the Vals Web Search Index. We built Keenable to learn from every agent query and it’s learning fast. artificialanalysis.ai/agents…

Post attachment
Artificial Analysis avatar
⟳ Wang shuai (Dylan) retweeted
Artificial Analysis@ArtificialAnlys·Nov 10

Total Cost per Task combines search cost and model inference cost. Keenable costs about $0.040 per task in both modes: about $0.032 for search (8 searches at $4 per 1k) and about $0.008 for the model. Search is about 80% of the Cost per Task. On Artificial Analysis Search Index vs. Cost per Task, Keenable (pro) is on the cost frontier: no product scores higher for less. Exa (lite, dynamic highlights) costs about the same and scores 79.

Post attachment
Artificial Analysis avatar
⟳ Wang shuai (Dylan) retweeted
Artificial Analysis@ArtificialAnlys·Nov 10

Keenable Search 1.0 takes 1st (pro, 83) and 2nd (realtime, 82) on the Artificial Analysis Search Index: ahead of Exa (auto, dynamic highlights) and Perplexity (medium) at 80 We benchmarked both modes of @KeenableAI's Keenable Search 1.0, pro and realtime, in our Stirrup agent harness with Keenable Search as its web_search tool to complete 1700 search-focused tasks. At $4 per 1k searches, Keenable costs about $0.040 per task in both modes, over a third below the board median ($0.063). It is also one of the fastest products we measure: 15.0s (realtime) and 15.5s (pro) per task, 2nd and 3rd on the board. Key benchmarking results for Keenable Search 1.0: ➤ Artificial Analysis Search Index: Keenable (pro) scores 83 and Keenable (realtime) 82, up from 67 for both modes when we first benchmarked Keenable in August, and a ~50-point lift over the same model with no search (33). Both lead Exa (auto, dynamic highlights) and Perplexity (medium) at 80 ➤ DeepSearchQA is Keenable's strongest benchmark: both modes score 88 F1, 1st and 2nd of 26, ahead of Exa (auto, dynamic highlights) at 86. Pro also ranks 1st on BrowseComp (87%, level with Perplexity (medium)) and on AA-Omniscience (73%, level with Exa (auto, dynamic highlights) and Firecrawl) ➤ Fast search: realtime takes 2.3s of search time per task (2nd of 26) and pro 3.0s (3rd), against a board median of 12.5s. Each search returns in about 330ms (realtime) and 390ms (pro) on average Key details: ➤ Setup: GPT-5.6 Luna (medium) in Stirrup, Keenable Search API with mode=realtime and mode=pro ➤ Pricing: $4 per 1k requests in both modes (public pay-as-you-go price) ➤ Cards: We disabled Keenable's search cards for testing (disable_cards=true)

Post attachment
Lenny Rachitsky avatar
⟳ Wang shuai (Dylan) retweeted
Lenny Rachitsky@lennysan·Nov 10

The best time to join a rocketship is before everyone knows about it. I’m excited to launch The Next 50: the 50 hottest early-stage startups to keep an eye on (and try to join). Using @harmonic_ai’s data, the juiciest startup dataset out there, we looked at every early-stage startup and ranked them on headcount growth, recent funding, and investor interest. We then picked the top 50. Here's the full list: lennysjobs.com/next50 We’ll update it quarterly, and I’d love your feedback on what other factors we should consider. P.S. This is in addition to the Lenny 100, the top 100 well-known companies with the highest talent density, tackling the most ambitious problems: lennysjobs.com/lenny100

Andrey Styskin avatar
⟳ Wang shuai (Dylan) retweeted
Andrey Styskin@styskin·Nov 6

I attended @GoogleStartups party yesterday and met many great founders. It was so cool. And 9 out of 10 of them need web search. Today, we’re launching Keenable’s AI Startup Program to support builders and give their agents access to the web. If you are backed by @conviction or @Accel, building something cool, or a friend, apply today:

Building an AI startup? Get $5,000 in Keenable credits for 6 months, plus onboarding and help promoting your launch. The fastest web search for agents: results in under 250 ms at p95 in US East. The credits work across every Keenable product. For startups backed by @Accel and @conviction, teams building something cool, and friends. Apply: keenable.ai/blog/ai-startups…