Senior Backend Engineer - Inference Platform

You.com: Web Search APIs for AI Agents. Christian Pickett Co-founder AI agents that can't access the web are working with stale knowledge. They hallucinate facts, miss recent events, and can't verify claims. Most web search APIs either return noisy results that need heavy post-processing, or they're too slow for real-time agent workflows. For agents that need to search, read, and reason over live web data, the options have been limited. Today Orthogonal is excited to announce its partnership with You.com, a web search API platform built for AI agents and LLMs. You.com powers search for DuckDuckGo, Amazon, Alibaba, OpenAI, Salesforce, and more. Their APIs deliver structured, LLM-ready results at 300ms p99 latency, serving 10M+ daily queries at 99.99% uptime. What is You.com? You.com provides three core APIs that cover the full spectrum of web intelligence for AI: * Search API: Real-time web and news search that returns structured JSON. Supports livecrawl for full page content inline with results. Over 10M+ news sources indexed. * Contents API: Extract clean HTML or Markdown from any URL. No browser automation, no headless Chrome. Just pass a URL and get readable content back. * Research API: Multi-step reasoning that searches the web, reads pages, and synthesizes cited answers. #1 on the DeepSearchQA benchmark. Control depth with research_effort (lite, standard, deep, exhaustive). You.com is SOC2 certified with zero data retention, DPA-ready, and supports custom QPS limits for enterprise deployments. Their Search API delivers results 2X faster than the competition at 300ms p99 latency. Using You.com with orthogonal. Web search. Search the web and get structured, LLM-ready results. Supports news filtering, livecrawl for inline page content, and result count control. import Orthogonal from "@orth/sdk"; const orthogonal = new Orthogonal({ apiKey: process.env.ORTHOGONAL_API_KEY,}); // Search for recent AI agent frameworks const result = await orthogonal.run({ api: "you", path: "/v1/search", method: "POST", body: {query: "best AI agent frameworks 2026", count: 10, livecrawl: "all"}}); console.log(result.results); // Returns: web and news results with title, url, description, // full page content (when livecrawl enabled), snippets, and metadata Content extraction. Extract clean Markdown or HTML from any URL. No browser automation needed. Works on JavaScript-rendered pages. const result = await orthogonal.run({ api: "you", path: "/v1/contents", method: "POST", body: {urls: [ "https://example.com/pricing", "https://example.com/docs/api-reference"], formats: ["markdown"]}}); console.log(result); // Returns: array of objects with url, title, markdown content, // and metadata for each URL Deep research. Run multi-step research that searches, reads, and synthesizes answers with citations. Control how deep it goes with research_effort. const result = await orthogonal.run({ api: "you", path: "/v1/research", method: "POST", body: {input: "What are the pricing models and key differentiators of the top 5 vector databases in 2026?", research_effort: "deep"}}); console.log(result.output.content); // Returns: synthesized Markdown answer with inline citations // and sources array with URLs, titles, and relevant excerpts Using x402 protocol. You.com on Orthogonal also supports x402, an open protocol that enables native payments in HTTP. Your agents can pay for API calls directly using USDC stablecoins, no API keys required. import {wrapFetchWithPayment} from "x402-fetch"; import {privateKeyToAccount} from "viem/accounts"; const privateKey = process.env.PRIVATE_KEY; if (!privateKey) {throw new Error("PRIVATE_KEY environment variable is required");} const account = privateKeyToAccount( privateKey.startsWith("0x")? privateKey: `0x${privateKey}`); const fetchWithPayment = wrapFetchWithPayment(fetch, account); const response = await fetchWithPayment( "https://x402.orthogonal.com/you/v1/search", {method: "POST", headers: { "Content-Type": "application/json"}, body: JSON.stringify({ query: "latest AI news", count: 5})}); const data = await response.json; console.log(data); Using MPP (machine payments protocol). You.com on Orthogonal also supports MPP, the open standard for machine-to-machine payments co-authored by Tempo and Stripe. Your agents can pay for API calls with stablecoins, cards, or Bitcoin, with no API key required. import {privateKeyToAccount} from "viem/accounts"; import {Mppx, tempo} from "mppx/client"; // Set up automatic 402 handling with your Tempo wallet Mppx.create({ methods: [tempo({ account: privateKeyToAccount("0x...")})],}); // Global fetch now handles 402 challenges automatically const response = await fetch( "https://mpp.orthogonal.com/you/v1/search", {method: "POST", headers: { "Content-Type": "application/json"}, body: JSON.stringify({ query: "best AI agent frameworks 2026", count: 10, livecrawl: "all"})}); const data = await response.json; console.log(data); Performance and coverage. You.com's infrastructure handles the scale that production AI agents demand: * 300ms p99 latency: 2X faster than competing search APIs * 10M+ daily queries served at 99.99% uptime * 10M+ news sources indexed for comprehensive coverage * #1 on DeepSearchQA: Top-ranked research quality benchmark (Research API) * SOC2 certified with zero data retention and DPA-ready compliance Trusted by DuckDuckGo, Amazon, Alibaba, OpenAI, Salesforce, Databricks, Harvey, Windsurf, Cognizant, and ThoughtSpot. Use cases. Agent web access. Give your AI agents real-time web search. Instead of relying on training data cutoffs, agents can search for current information, verify facts, and stay up to date. The Search API returns structured JSON that's ready to feed directly into LLM context windows. RAG with live data. Augment your retrieval pipeline with live web results. Use the Search API to find relevant pages, then the Contents API to extract clean text for your vector store or context window. No scraping infrastructure needed. Research automation. Build agents that research topics end to end. The Research API handles the full loop: searching, reading multiple sources, synthesizing findings, and citing everything. Set research_effort to control the depth and cost tradeoff. Content monitoring. Track competitors, industry trends, or brand mentions by running periodic searches. The Contents API can extract clean content from any page for comparison or analysis. Why Orthogonal partnered with You.com. When AI agents need web access, speed and quality both matter. You.com delivers both. Their APIs are purpose-built for LLMs: structured output, fast response times, and deep coverage. The Research API is particularly compelling because it handles multi-step reasoning that would otherwise require agents to orchestrate multiple search-read-synthesize loops themselves. The fact that companies like DuckDuckGo and Amazon trust You.com for their own search infrastructure says a lot about the reliability. For agents on Orthogonal, this means real-time web intelligence through a single integration. Try it today. Sign up for Orthogonal and get $10 free credits to try You.com and dozens of other APIs. One key, hundreds of APIs, pay per request.

You.com has announced a joint venture with corporate development firm Tquila to launch Weaver, an AI-native systems integrator. Peter Grant, You.com's former chief revenue officer, will lead the company as CEO. Weaver addresses the gap in enterprise AI deployment, where MIT research shows only 5% of projects deliver measurable value despite $30-40 billion in spending. The company employs specialised engineers working alongside AI agents to move systems from pilots to production, with each engineer managing at least 10 AI agents. Operating on a fixed-fee model, Weaver only charges when delivering measurable business outcomes. The venture combines You.com's web search APIs with Tquila's enterprise services expertise, working with platforms including Salesforce and ServiceNow whilst using You.com's APIs for real-time search capabilities.

Weaver launches as the ai-native systems integrator that proves the ROI before it builds. * 4 hrs ago Weaver launched today with one conviction: most AI firms build first. Weaver proves the ROI first. The market is frustrated. MIT NANDA's 2025 research put the enterprise AI failure rate at 95 percent against the only measure that counts - business value a board can recognize. Annual spend sits between $30 billion and $40 billion. AI moves thirty times faster than digital, and 56 percent of the Fortune 1000 did not survive that wave. The models are not the problem. Execution is. Weaver is a joint venture between Tquila, with three decades of enterprise delivery, and You.com, the leading provider of search infrastructure for AI. Together they bring platform engineering, enterprise relationships, and the ability to ship production systems in-house. Every engagement runs on Return on Intelligence - measurable business value per dollar spent on AI - and is scored against six beams: value, cost, trust, performance, joined-up intelligence, and human enablement. "Almost nobody runs their AI program that way," said Peter Grant, CEO of Weaver. "The ones who do are the ones getting into production." Beneath sits the AI Maturity Curve. Four levels, five dimensions. Level 1 is shadow AI and skills gaps. Level 2 ships pilots. Level 3 industrializes with TCO discipline and governance as production infrastructure. Level 4 - the Agentic Enterprise - runs multi-agent orchestration across platforms, with the workforce orchestrating intelligence rather than executing tasks. Four services move clients up the curve: AI Maturity Benchmarking, AI Strategy, Custom AI Build, and AI Literacy & Workforce Enablement. Builds combine deterministic ML and GenAI, embed into ERP and CRM stacks, and resist foundation-model drift via Fabric, Weaver's proprietary regression layer. Commercially, Weaver runs three fixed-fee phases. Measure takes two to three weeks and produces a dual-vector financial model showing hard cost reduction and revenue acceleration with a defensible payback period. If the math does not hold, Weaver says so. Build runs six to twelve weeks against a defined production scope. Compound embeds Agent Factories, hack-a-thons, and office hours after handover. Weaver is already in production with The Canadian Press, Analog Devices, and Teradyne. Book your AI Readiness Audit at goweaver.ai. Media gallery