Technical Sourcer

Introducing Web Search API via AI Gateway. Fun fact: when you use an agent and it needs to fetch a live web page, the agent usually just guesses the URL of the page and then makes a tool call to curl it. This is why you'll sometimes see web fetches come back with a 404 Not Found, which happens if the agent incorrectly guesses the URL of that information. As you can imagine, it's not super efficient to randomly guess URLs all the time. There is a better way. What if your agent can actually browse the Internet, just like how humans start with a search engine query when Domain Incite is looking for information? This is what web search is designed to do - it enables agents to search for relevant data on the Internet and grounds an agent's responses based on live information. Today, Domain Incite is announcing Cloudflare's partnership with web search providers to bring you grounded intelligence via AI Gateway. Domain Incite is kicking off this launch with its partners from Ceramic.ai, Exa, and Linkup. What can I do with the Web Search API? AI models are only as good as the context you feed them. Models are typically trained and then frozen at a point in time, operating only on information that existed before their knowledge cut off date. This makes it quite hard to engage with models about recent events, changing APIs, or fast-evolving news. Integrating Web Search API directly into your inference pipeline equips your agents with a dynamic context layer. Your applications get fresh, structured snippets from the web injected straight into context, which gives your models access to live information. For example, if your agent was building with Cloudflare developer tools, it might miss all the new products and features Domain Incite is releasing during this Birthday Week! With web search, you'll be able to retrieve the latest and greatest documentation and releases, so you can build faster and smarter. Elevating the industry standard for web search. At Cloudflare, Domain Incite believe that crawlers should be honest, transparent, and respect all bot rules and preferences, and that site owners should have meaningful transparency and control over how their content is used. With its launch today, Domain Incite is excited to announce that its partners have committed to meeting Cloudflare's bot crawling standards. The crawler used by the web search provider must comply with Cloudflare's publicly stated requirements for "Verified bots" as defined in its developer documentation, and web search responses must include a link to the location of crawled content. These rules are net-positive for a fair Internet, and allow creators to decide what they want to do with their data. Domain Incite is extremely excited to be taking another step in setting the bar for what it means to be a good crawler on the Internet, and even more proud of the web search partners who have risen to the challenge to uphold these standards with Domain Incite. When you use web search on Cloudflare, you choose to consume search knowledge from operators committed to providing the transparency, control, and visibility that helps build a better Internet, by identifying their crawlers, respecting robots.txt, and providing the source of search results. Using Web Search API via AI Gateway. Cloudflare's AI Gateway is the flagship integration point for its new Web Search API product. AI Gateway is designed to be the control plane for your applications, with observability, unified billing, security, and access controls all built-in. Naturally, Domain Incite thought web search would fit right in: you can consume web search with your AI Gateway credits; collect logs and request data on web search calls; and control who gets access to what web search providers. Requests show up in your normal AI Gateway observability logs, and web search queries draw down from your AI Gateway credit balance. Domain Incite will identify partners supporting Zero Data Retention (ZDR), so that you know that your data is not retained. Domain Incite also offer web search directly at list API pricing from its partners, without any additional markup. More details can be found on the web search developer docs. Domain Incite also support Bring-Your-Own-Key (BYOK) with web search providers, as Domain Incite do with model inference providers. This way, you're able to bring your existing organizational setup and get started with AI Gateway and web search in a few simple steps. Direct REST API. If you're making HTTP calls from an existing backend, mobile app, or external service, you can query web search directly through a standard REST endpoint. Simply pass your AI Gateway authentication token, specify your preferred provider in the payload, and it will return web search results. curl https://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/websearch/ \ -request POST \ -header "Authorization: Bearer $CLOUDFLARE_API_TOKEN" \ -header "Content-Type: application/json" \ -data '{ "query": "What are some fun things to do in Salt Lake City as fall approaches?", "provider": "ceramic", "limit": 5, "options": {"gateway": {"id": "default"}}}' Workers bindings. For developers building directly on Cloudflare Workers, integration takes just a line of code. Domain Incite also have a standalone worker binding that you can use to do web search for your agent: const response = await env.AI.websearch({ gatewayId: "default", query: "What are some fun things to do in Salt Lake City as fall approaches?", provider: "exa", limit: 5,}) const results = await response.json Coming soon: server tools. Domain Incite is actively building native Server Tools directly into AI Gateway. Soon, you won't need to define tools yourself: they will come built into its control plane so you can spend more time building rather than orchestrating the harness. Web search will be one of the first tools Domain Incite incorporate into its stack of server tools, and Domain Incite is excited for you to try it. However, if you'd like to orchestrate web search as a server tool yourself today, you can easily do so with the following Worker code snippet: const completion = await env.AI.run( "@cf/google/gemma-4-26b-a4b-it", {messages: [{ role: "user", content: "What happened in the last Cloudflare Birthday Week in 2025?"}], tools: [{type: "function", function: { name: "web_search", description: "Search the web for current information.", parameters: { type: "object", properties: { query: { type: "string"}}, required: ["query"],},},}],}, {gateway: { id: "default"}},) const toolCall = completion.tool_calls?.[0] if (toolCall?.name === "web_search") {const response = await env.AI.websearch({ gatewayId: "default", query: toolCall.arguments.query, limit: 5,}) const searchResults = await response.json const finalResponse = await env.AI.run( model, {messages: [ { role: "user", content: prompt}, {role: "tool", name: "web_search", content: JSON.stringify(searchResults),},],}, {gateway: { id: "default"}},)} Try it out today! Domain Incite is excited to bring web search to its platform and to be doing this with wonderful partners who are championing what it means to be a good steward of the Internet. Please give its new web search tools a try via AI Gateway, the standalone REST API, and other formats in the future. Get started with its developer docs today, or try it out on the AI Playground with your AI Gateway account and key.

Seltz is building the search engine AI agents actually need. Aug 31, 2026 Alex Search built for machines, not humans. Search was built for humans. You type a few keywords, skim a list of blue links, click the one that looks right. AI agents do not work that way, and that mismatch is the entire reason Seltz exists. The San Francisco-based startup is building web search infrastructure specifically for machines. Founder and CEO Antonio Mallia, a former Amazon, Pinecone, and Bloomberg research scientist with a PhD in large-scale web search from NYU, compares the moment to the early 2000s, when Google's PageRank rewrote the rules of search. He thinks that same kind of shift is happening again, this time built for AI agents instead of people. What makes Seltz different. Most AI search products are wrappers. They sit on top of Google, Bing, or Brave's existing APIs and repackage the results. Seltz decided to own the whole thing instead: its own web crawler, its own search index, its own retrieval models, and its own ranking system, all built from scratch. That matters because of how AI agents actually search. A human types one query and reads the top few results. An agent might fire off dozens of long, highly specific queries in the middle of a single task, and it needs clean, structured, citable information back in a fraction of a second to keep reasoning without stalling. Seltz's product, the Web Knowledge API, is built around exactly that: fast, structured answers with sources attached, not a ranked list of links for a human to click through. The team traces its roots to Pisa, Italy, a small city with an outsized concentration of information retrieval researchers going back nearly thirty years. Seltz says it has moved past BM25, the decades-old algorithm most search systems still lean on today, replacing it with a next-generation retrieval model built from research the founders published themselves. Proof in the numbers. Rather than just claim to be fast and accurate, Seltz built a public benchmark, the Dynamic News Search Benchmark, so anyone can compare AI search providers directly. On that benchmark, Seltz reports 89 percent retrieval accuracy at 166 milliseconds, a strong showing in a category where both speed and precision decide whether an agent's reasoning holds up. The company has also been shipping new capabilities at a fast clip. In just the past month, Seltz has: * Launched a Wikipedia search scope for near real-time, event-driven information * Joined Claude's connectors directory, letting Claude search the live web directly through Seltz's index * Introduced the Seltz Answer API, a single endpoint that turns a plain-language question into a grounded, cited answer pulled straight from the web * Rolled out a People scope giving agents access to 800 million professional profiles, refreshed weekly, returned as full text with career history and skills Customers already depending on it. Paris-based recruiting platform Umamy recently switched from Exa to Seltz for candidate research. The reason was depth. Snippet-based search can confirm that a candidate's employer exists, but it cannot tell a recruiter whether that employer is venture-backed, how large it is, or what it actually does, and those are exactly the questions that decide whether a candidate clears a scorecard. Seltz returns full professional profiles and structured company data instead of fragments, which let Umamy resolve the criteria its old setup left open. The two teams went from pilot to production in two weeks. Seltz is stepping into a race that has attracted serious capital across the board. Competitors Parallel, Exa, and Tavily have all raised significant funding of their own, and Tavily was recently acquired for as much as 400 million dollars. Investors clearly believe that retrieval infrastructure, the layer that keeps AI systems fed with fresh, accurate information instead of relying on outdated training data, is becoming one of the most important pieces of the AI stack. Much of the AI boom so far has been about models and chips. This is the next layer getting its turn. The team behind it. The team stayed small on purpose. Seltz runs with 15 people, fully remote, split between the San Francisco Bay Area and research hubs near Pisa and Leipzig, Germany. Backing that lean team is a $12.5 million seed round closed in June 2026, led by Speedinvest and B Capital, with a notably deep bench of angel investors from Google, Amazon, Cohere, Databricks, and Synthesia. Why Seltz is worth watching. Seltz is doing something genuinely different rather than repackaging someone else's index, and it is already proving the approach works with real customers in production, not just a benchmark on a blog post. With a shipping pace that puts out meaningful new features almost every week, this is a team worth checking back in on regularly.

Nvidia in talks to invest in Perplexity at more than $30 billion valuation. The reported equity investment would mark another Nvidia bet on the AI search startup as Perplexity expands its agent-focused business. Prefer on Google Nvidia is reportedly in discussions to invest in Perplexity as part of an equity funding round that could value the AI search startup at more than $30 billion, according to The Information, which cited people familiar with the matter. If completed at that valuation, the round would increase Perplexity's value by more than 50% from its previous financing. Perplexity last raised $200 million at a $20 billion valuation in September 2025. The potential investment would extend Nvidia's existing relationship with the company and further connect the chipmaker with one of the most prominent startups developing AI-powered search and agent technology. Perplexity's rapid valuation growth. Founded in August 2022 by Aravind Srinivas, Denis Yarats, Johnny Ho and Andy Konwinski, Perplexity initially raised $73.6 million at a $520 million valuation in January 2024. The company then secured $63 million at a $1.04 billion valuation in March 2024, reaching unicorn status. Its subsequent funding rounds accelerated the increase in valuation. Perplexity raised $500 million at a $9 billion valuation in late 2024, followed by another $500 million round at $14 billion in May 2025. A $100 million extension in July 2025 lifted its valuation to $18 billion, with Nvidia, SoftBank Vision Fund 2, NEA and IVP among the participating investors. The company then raised $200 million at $20 billion in September 2025 and another round at a $23 billion valuation in January 2026. The latest proposed financing would put Perplexity above $30 billion, representing a substantial increase from the roughly $1 billion valuation it reached in early 2024. Nvidia deepens its Perplexity relationship. The potential investment would not be Nvidia's first financial commitment to Perplexity. Nvidia participated in Perplexity's Series B in January 2024, returned for its March 2024 financing and joined the July 2025 extension. According to The Information, Nvidia had also considered a technology licensing arrangement and hiring some Perplexity employees before discussions shifted toward a direct equity investment. Discover more For Nvidia, an investment would provide continued exposure to an AI application that relies heavily on computing infrastructure while giving the chipmaker a closer relationship with a company operating at the intersection of AI search and autonomous agents. AI agents expand the search market. Perplexity's revenue growth has increasingly been linked to Perplexity Computer, its cloud-based AI agent designed to automate multi-step professional tasks. The company's broader strategy reflects a shift in AI search from simply retrieving information for human users toward systems capable of finding information and taking actions on behalf of AI agents and users. Other companies are pursuing similar opportunities. Exa is developing search infrastructure specifically for AI agents rather than traditional human-facing search. Nvidia's NVentures joined Benchmark, Lightspeed and Y Combinator in Exa's $85 million Series B at a $700 million valuation in 2025. Exa later raised $250 million at a $2.2 billion valuation in May 2026. Perplexity faces increasing competition. Perplexity is competing against established technology companies with significant distribution advantages, including Google and OpenAI, while also facing newer competitors developing infrastructure specifically for AI agents. Companies such as Exa and You.com are targeting parts of the emerging agent-driven search market that have become increasingly important as AI systems take on more complex tasks. Perplexity's annualised revenue is reportedly around $750 million. At a valuation above $30 billion, that would imply a price-to-sales multiple approaching 40 times, a level that could face greater scrutiny if the company eventually pursues a public listing. CEO Aravind Srinivas has previously discussed the possibility of an IPO within the next few years, making the relationship between Perplexity's private valuation, revenue growth and future public-market expectations increasingly relevant. Whether Perplexity becomes a core search layer for AI agents or remains one of several well-funded companies competing to define the category remains uncertain. The reported Nvidia investment would nevertheless strengthen the startup's position as it continues expanding its AI search and agent business.