A Y Combinator startup called Click went live on August 12, 2026 with a tool that plugs LinkedIn profiles, X feeds, YouTube channels, and corporate financial filings straight into ChatGPT and Claude. The pitch is simple: AI assistants have gotten good at conversation but remain bad at actually finding things.
Founder Aditya Asgaonkar, who spent years working on Ethereum consensus protocols before this, describes the problem on the company's site as "the clearest gap today" in how people use AI. Foundation models can search the web, sure, but they stumble when asked to pull structured data from platforms that don't surface well in standard search results. Click's connectors are designed to fix that, acting as a bridge between the chat interface and the APIs most users can't easily reach on their own.
The product runs as what's called an MCP server. That's Model Context Protocol, an open standard Anthropic released to the Linux Foundation earlier this year. Once installed, it lives inside the AI assistant itself. Users get access to connectors for LinkedIn, X, YouTube, Reddit, Google Maps, Zillow, travel booking sites, and SEC filings, according to the company's feature list.
Click is targeting three types of work: sales prospecting, event planning, and influencer research. The examples on its homepage suggest the kind of scale it's aiming for. One workflow shows the system reviewing three LinkedIn profiles, 24 posts, and 43 reactions to build a prospect dossier. Another evaluates 214 flights and 86 hotels while simultaneously parsing earnings reports for conference logistics. A third analyzes 72 X accounts and 14 YouTube creators for campaign research.
"I live inside their surface now," Asgaonkar wrote on Product Hunt, referring to Claude and ChatGPT. "I built the connector layer that I wish existed."
The Distribution Play
The product works across Claude's web, desktop, and API versions through its connector menu. For ChatGPT, users need to enable Developer Mode with MCP support, per documentation from both Anthropic and OpenAI. Click charges $49 a month for individuals, which includes 500 credits and automatic access to new connectors as the company ships them.
It's a pricing model that assumes people will pay to collapse the friction between asking an AI a question and actually getting a useful answer. Whether that holds depends on how often users hit the limits of built-in search, and whether $49 feels reasonable for solving that problem.
Asgaonkar's background tilts technical. Before Click, he worked as a researcher at the Ethereum Foundation, contributing to consensus protocol development. His published work includes papers on fast confirmation rules and formal verification of the beacon chain, according to arXiv and Springer listings. "Chat is becoming a place to do work," he said on the company's site. "Research is Click's first service."
Click AI Labs incorporated in Delaware and shows one employee on LinkedIn as of late August. It participated in Y Combinator's Summer 2026 batch and joined NVIDIA's Inception Program, though the company declined to share funding figures or valuation.
A Crowded Moment for Research Connectors
Click isn't alone in this. The past four months have seen a small wave of MCP-based research tools. Scite launched a citation-backed connector for Claude in April. Dimensions Analytics went live with scholarly data access on August 11, one day before Click. Versive, Scale Insights, and Sprig all released MCP integrations for user research, Amazon sales analytics, and survey data between May and August.
The common thread: AI assistants now handle multi-step research tasks, but the built-in web search they rely on doesn't reach platform APIs or data sitting behind login walls. A blog post from Bigdata.com in May illustrated the gap with a test showing Claude producing different financial analysis depending on whether it used generic web search or a dedicated financial-data MCP. That's the territory Click and its peers are staking out.
Early traction looks modest. Click's Product Hunt launch netted 134 upvotes and a ninth-place daily ranking. The company's LinkedIn page had 373 followers as of August 21–22. Product Hunt reviews stood at zero around the same time, though that's not unusual for B2B tools that take time to evaluate.
"We're live!" the company posted on LinkedIn in mid-August. "Install the Click MCP in ChatGPT/Codex or Claude and give it a try!"
There's a privacy note worth flagging. Click's policy, updated August 3, advises users not to submit highly sensitive information. The service logs MCP requests for operational purposes and may share data with connector providers when a user's AI agent requests a specific tool. That's standard for this kind of middleware, though it's something enterprise customers will scrutinize.
The broader question is whether research connectors become a feature or a product. If Anthropic and OpenAI decide to build deeper platform integrations themselves, startups like Click could find their niche squeezed. For now, though, the gap exists, and Click is betting there's a business in filling it.
