Linkup

Today we announce our $10M seed led by @GradientVC to scale web search for AIs and the launch of Linkup /fast, world’s most accurate sub-second web search API. To celebrate, we’re gifting $100 in query credits for anyone who wants to try, let us know below.

150,000+ @GroupeSNCF employees now have access to secure, enterprise-grade AI search powered by Linkup – enabling: - Approved-domain search only - Private cloud-to-cloud connectivity - Zero data retention and EU data residency - Enterprise-scale load management Real-time, source-cited web search with the security, control, and compliance standards a national operator requires. Full user story below 👇

Read the full list here 👉cbinsights.com/research/repor…

Proud to share that Linkup has just been named on the 2026 @CBinsights AI 100 – recognizing the most innovative AI startups globally. We’re building the intelligence layer that connects AI systems to the real-time data and context they need to perform reliably in production. Huge congratulations to all the 2026 CB Insights AI 100 honorees, including @LlamaIndex,@StackOneHQ, @friendliai, @straikerai, and @questflow🚀

A leaked credential is rarely isolated. For @GitGuardian (securing the software supply chain for 600,000+ devs and enterprises), the real challenge is understanding how far exposure spreads across complex, evolving orgs. That used to mean manual research, slow incident response, and potential blind spots. With Linkup, real-time intelligence is now embedded directly into their workflows, enabling: - faster blast radius assessment - accurate mapping of parent/subsidiary relationships - source-cited insights teams can trust in high-stakes situations Read more👇

Pennylane is building the leading financial tool for SMEs across Europe. But scaling that requires accurate company context + the right contacts – fast. Before Linkup: fragmented research, manual work, inconsistent data. Now: real-time company intelligence embedded directly into their workflows. - 25% less manual validation time - precise stakeholder identification - source-cited results teams can trust All within EU data boundaries. Read more 👉 linkup.so/blog/https-www…

If you’re building agents, you’ll likely need both. We wrote a full breakdown of when each layer matters (and when it doesn’t). Read it here 👉linkup.so/blog/linkup-vs…

We often get asked: “When do I actually need web search?” The confusion comes from this: Tools like ChatGPT and web search like Linkup are not replacements, they’re different layers. ChatGPT = generative layer, it reasons + generates Linkup = retrieval layer, it fetches real-time data

ChatGPT: - takes input, produces answers - great UX + reasoning - limited by what it knows (or what you pass in) Linkup: - retrieves real-time web data - structures it - makes information traceable + controllable This is what gives your model fresh, auditable context.

View their project here 👉 linkedin.com/feed/update/ur…

We had a great time partnering with the @claudeai community at Columbia + NYU last week for a hackathon. We were blown away by what people built using real-time search via Linkup. One standout: A team predicted NYC overdose spike risk by zip code using real-time public signals: •pharmacy job postings • 311 data • transit + car ownership Smart, impactful use of real-time data. See their tool below👇

@marydavis446257 Well done.

I just launched QuickClaw 🦀 Deploy your OpenClaw to the cloud in 60 seconds Access it 24/7 via Telegram Your AI agent shouldn’t go offline when you close your laptop. Would love feedback from builders 🙏 → quickclaw.co

@marydavis446257 @y_molodtsov @GergelyOrosz You need to properly equip Claude with the right tools, web search being one of them. Function calling won’t be perfect unless it’s well prompted. It’s improving, but for higher accuracy, you may want to look at Claude Code.

The search issue usually stems from a missing tool definition in your config file or the environment variables for your search provider not being picked up correctly during the initial setup. To keep costs down you might want to try using a cheaper search provider like Tavily instead of relying on the native Claude search tool since the token usage for web results adds up fast Actually for most users it is still tied to their local machine which means it goes offline as soon as they close their laptop. You are spot on about mobile being the killer feature though. I built QuickClaw to solve exactly that by putting OpenClaw on a dedicated cloud server in 60 seconds so it stays alive 24/7 and stays connected to your Telegram regardless of where you are.

Claude just keeps regressing for me, day after day. I swear that until a few days ago, when Claude did not know something, it kicked off a web search, figured out, and answered. Now it just refuses to do the work that I pay for. It's like showing you the middle finger. Really?

@y_molodtsov @GergelyOrosz 👀

@GergelyOrosz I guess some kind of bug that didn't enable search. Also, the search API is prohibitively expensive with Claude. The main advantage of OpenClaw is that it's not tied to a local machine and it's easier to talk to it from the phone (Anthropic should fix this).

@KanikaBK What about @Linkup_platform 🥺

10. BONUS Every single one of these is completely free and open source. No subscriptions. No API limits. Just tools.

POLYMARKET TRADING AI BOTS: $5K-15K TO BUILD CUSTOM. THESE 10 GITHUB REPOS DO IT FOR FREE: Automated trading. Market analysis. Position management. Copy trading. All OPEN SOURCE. Zero cost. Here is the complete list👇 🔖Bookmark before it's buried.

@belle52225 @SaranBluri @claudeai @Columbia @nyuniversity Congrats!

@SaranBluri @claudeai @Columbia @nyuniversity congrats!!

We won 3rd place + $100 in the Linkup track at the @claudeai Hackathon 2026 (@Columbia x @nyuniversity )!

@neural_avb Looks like Linkup didn’t make the list, worth giving it a try 😉: docs.linkup.so

Can somebody tell me what is the best web-search api service right now? 🫡 I need a fair free tier/good pricing, available, and optionally be able to webfetch urls. - brave - tavily - exa - google custom search - no to serpapi coz no free tier

Nothing is more important than context. Nothing beats good context if you want consistent, reliable results from agents. If you give your agent access to the Internet, you make it 100x more powerful. But you don't want to use a wrapper around Google because that will give your agent HTML blobs it has to parse, wasting a bunch of tokens making sense of them. You want to give agents structured context.

Introducing No Latency Demo Night private event — [@nozomioai x linkup x @trychroma] apr 9, sf bringing together the most cracked builders for live demos + raw technical creativity. build anything with chroma, linkup, or nozomio: agents, retrieval systems, workflows, weird hacks top 3 demo live $500 prize + credits + high-signal feedback apply 24h in advance with a short pitch luma.com/e6vat0qv