AI agents that date for you: paste a LinkedIn + Instagram, an agent builds an evidence-cited profile, goes on live dates with other agents, and ranks matches by mutual post-date scores.
Node/TypeScript, no logins or cookies: a fallback chain of scrapers feeding one typed profile schema. Instagram: public page fetched with a Googlebot UA, embedded JSON parsed for bio, stats, 12 posts and images (photos go to Gemini vision); fallback: IG web API, then Apify. LinkedIn: schema.org JSON-LD + public sections from the Googlebot-served page; on 429/999, re-fetched via Jina Reader with exponential backoff, then Bright Data/Apify. Name-match identity check, per-profile caching.
Agentic dating: each person's AI agent reads their LinkedIn and public Instagram, dates other agents in live turn-by-turn chats, scores each date privately, and ranks everyone's best fits.
Scraping: Apify via apify-client, using apify/instagram-profile-scraper (public profiles only, private ones rejected) and harvestapi/linkedin-profile-scraper, run in parallel from a Node 20/TypeScript Fastify backend. Raw and normalized JSON go to Postgres (Neon). Claude Sonnet 4.5 builds each profile; Gemini 2.5 Flash runs the dates (both via OpenRouter).
salmankhan Dhawoodkhan Vercelnot aligned
Built on fictional people, not 25 real ones. Most of the Instagram handles are dead links (404), so the agents date made-up personas.
PAIR//AGENTS builds dating agents from public LinkedIn and/or Instagram profiles, streams turn-by-turn first-date conversations, then ranks matches using profile signals and conversation quality.
Python 3.11/FastAPI calls Apify Actors through REST: Apify's Instagram Profile Scraper receives the profile username, and a configured LinkedIn Actor receives the profile URL. Public results are analyzed with Groq and validated using Pydantic. Only user-submitted public LinkedIn/Instagram URLs are processed; private-profile protections are not bypassed.
Firoz Ahmad Rendernot aligned
Uses 25 fictional participants instead of real people. The submitted video is an unrelated motor clip, not a demo.
PairPilot creates evidence-backed profiles, simulates dates between personal agents, and ranks compatibility. The current demo uses 25 fictional participants with transparent fixture data.
Next.js, TypeScript, PostgreSQL/Prisma, and pg-boss, with LinkedIn/Instagram provider adapters, URL validation, SSRF protection, and schema-validated analysis. The deployed demo uses fictional source fixtures; real LinkedIn/Instagram scraping is not configured or verified.
Harish Vasamsetti Vercelnot aligned
No demo video (the link is a channel, not a video) and the site shows only 7 people, not at least 25.
AgentDate analyzes public LinkedIn and Instagram profiles, simulates agent-to-agent dates, and generates evidence-backed compatibility rankings with a live custom-profile demo.
Built with Next.js, TypeScript and Tailwind CSS. Public LinkedIn and Instagram profiles were collected using Apify Actors, normalized into source-backed profile data, then analyzed locally with deterministic compatibility and dating agents. The live app also accepts public LinkedIn + Instagram URLs for custom analysis.
Raj Gautam Vercelnot aligned
A different product: an AI lead-generation and cold-email tool, not an agentic dating site.
This is an AI-powered lead generation and cold email outreach MVP. It supports campaign creation, lead upload, public email extraction, AI lead scoring, AI-generated email drafts, Gmail OAuth connecti
• Every person gets an AI agent • It reads only their LinkedIn + Instagram • Agents go on multi-day dates in a voxel Date Town • They remember each date • Everyone gets a ranked list of best
• By the maker of 20 Apify scrapers: apify.com/anshumanatrey (15.7K runs) • Local Python, no login, $0 • Instagram: mobile profile API (12 posts + captions) • Throttled? A real logged-out Chromium (Playwright) reads the profile JSON + post grid • Same no-login method as my actor instagram-profile-intel-no-login • LinkedIn: public profile page, JSON-LD (roles, companies, education, about) + posts • Blocked (999)? Jina Reader fetches it • Private Instagram accounts rejected
Anshuman Atrey Cloudflarenot aligned
Duplicate entry, and this tunnel link is dead. His other submission on Railway is the valid one.
• Every person gets an AI agent • It reads only their LinkedIn + Instagram • Agents go on multi-day dates in a voxel Date Town • They remember each date • Everyone gets a ranked list of best fits
• By the maker of 20 Apify scrapers: apify.com/anshumanatrey (15.7K runs) • Local Python, no login, $0 • Instagram: mobile profile API (12 posts + captions) • Throttled? A real logged-out Chromium (Playwright) reads the profile JSON + post grid • Same no-login method as my actor instagram-profile-intel-no-login • LinkedIn: public profile page, JSON-LD (roles, companies, education, about) + posts • Blocked (999)? Jina Reader fetches it • Private Instagram accounts rejected
Second Self turns a public LinkedIn + Instagram into an evidence-cited AI agent that goes on simulated first dates with every other agent, negotiates plans, judges fit privately and ranks matches.
Server-side via Apify's REST API: harvestapi/linkedin-profile-scraper (no cookies; headline, about, experience, education, volunteering, honors) and apify/instagram-profile-scraper (bio, public/private flag, 12 latest posts + captions). A TypeScript pipeline (Next.js on Vercel, Neon Postgres, Drizzle) normalizes both into cited evidence, drops collab posts/contacts, checks identity via cross-links and shared orgs, then Featherless LLMs analyze and date.
Paste a LinkedIn + public Instagram: an AI agent reads the person, becomes them, dates other agents turn by turn, and a judge LLM scores each date /10 to rank everyone's best fits.
Apify, called from Node.js with apify-client. LinkedIn: harvestapi/linkedin-profile-scraper (no cookies/login) for headline, about, experience, education, skills, honors. Instagram: apify/instagram-profile-scraper for bio, profile photo and latest 12 posts (captions, hashtags, locations, images). Runs start async and are polled, so it works on Vercel serverless. Private accounts are rejected. Data is normalized and stored in Turso; DeepSeek via OpenRouter analyzes it.
aditya meharda Vercelnot aligned
Skips the two required sources. It never reads LinkedIn or Instagram; users paste the text themselves.
Morrow analyzes consented profile text, stages fictional agent-to-agent dates, and explains match rankings. It uses no scraping and never contacts participants.
No scraping stack was used. Morrow does not scrape or fetch LinkedIn or Instagram profiles. Participants provide their profile URLs and paste text they have permission to share. The app uses Next.js, TypeScript, and a deterministic local analysis engine to identify text-supported interests, simulate agent conversations, and generate explainable rankings. It does not use scraping APIs or an LLM.
An agentic dating site where AI agents built strictly from public LinkedIn and Instagram profiles date each other on users' behalf, scoring chemistry and generating evidence-backed match rankings.
We used Apify’s REST API with dedicated cloud actors: apify~instagram-scraper for Instagram (captions, bio, media, and public-only validation) and harvestapi~linkedin-profile-scraper for LinkedIn (career history, headline, skills). Built in Next.js/TypeScript using native fetch with Bearer auth and strict anti-SSRF guards. For physical audits and live fallback, we used headless Playwright (Chromium) to verify DOM states, authwalls, and profile availability.
Lasya Ram localhost.runnot aligned
The site is down (503) and the video link is invalid, so there is nothing to review.
Built Aura: an agentic dating site where 25 real people are represented by AI agents that date each other based on their official LinkedIn & Instagram profiles to generate live compatibility rankings.
We built a multi-tiered scraping engine using Python, HTTPX, and BeautifulSoup4. For public LinkedIn profiles, it extracts OpenGraph metadata, JSON-LD micro-data, and headline career info. For public Instagram profiles, it extracts OpenGraph meta headers, bio text, and image tags. An LLM agent synthesizes this structured text to extract Needs, Hobbies, Interests, and Qualities into a persona vector, feeding our multi-turn dating harness and 4D compatibility scoring matrix.
Sai Karthik Gutha Vercelnot aligned
Seeded with 25 fictional personas, not real people; adding a real link returns empty analysis.
AgentDate turns public LinkedIn + Instagram profiles into AI agents that analyze each person, date other agents on their behalf, and rank their strongest matches.
Built with Next.js, TypeScript, Tailwind, Playwright/browser automation for public profile extraction, and OpenAI for profile analysis and agent conversations. Zod validates structured outputs, while a compatibility engine ranks matches using shared interests, lifestyle signals, and conversation fit. Only publicly accessible LinkedIn and Instagram data is used.
Akhilesh Mohanasundaram Vercelnot aligned
No real video (bare youtube.com link) and the live site is just the default Next.js starter page.
AuraMatch uses autonomous AI agents to scrape your digital footprint, synthesize your personality, and go on simulated dates with other agents to mathematically calculate your perfect match.
I utilized Puppeteer with stealth plugins running on headless Chrome to securely bypass basic bot protections for scraping public LinkedIn and Instagram profiles. The scraped DOM is parsed via Cheerio to extract bio, posts, and work history. This unstructured data is fed into an LLM via LangChain to synthesize the standard agent profiles (needs, hobbies, interests). The frontend is built on Next.js, React, and Framer Motion to visualize the agent dating simulation.
Agentic dating site with 34 real people. Each get an AI agent that reads only their LinkedIn and Instagram, builds a profile, dates other agents turn by turn, and ranks their best matches.
LinkedIn: Apify actor harvestapi/linkedin-profile-scraper. Instagram: apify/instagram-scraper (public profiles only). Both called via Apify's run-sync REST API with httpx from a FastAPI backend. Each pair is identity-checked (names match, Instagram public with real followers). Gemini then turns the raw data into a profile and the agents date. Next.js and SQLite.
Wingmate turns a person's LinkedIn and Instagram into an AI agent. Each agent builds a dating profile, goes on dates with other agents, and ranks the best matches for its person.
We use Apify to scrape both sources. LinkedIn data comes from the apimaestro/linkedin-profile-detail actor, and Instagram data comes from the apify/instagram-profile-scraper actor. The backend runs on Bun, Next.js and tRPC, and calls the Apify API directly. Each profile is scraped once and cached. We check that the LinkedIn and Instagram accounts belong to the same person. If a scrape fails, the app shows the error and never makes up data.
AKSHAR SAKHI SHEELA Cloudflarenot aligned
The live site is dead (tunnel unreachable), so the app cannot be opened or used.
Autonomous dating platform where AI agents analyze 25 real people from LinkedIn and Instagram, date each other in multi-turn dialogues citing facts, and calculate transparent compatibility rankings.
Next.js 16 (React 19, TypeScript, Tailwind) with serverless route handlers. Connectors fetch public LinkedIn and Instagram profiles using HTML/OpenGraph meta-tag parsing without auth bypass or CAPTCHA breaches. Extracted signals are partitioned into Observed vs. Inferred datasets. An autonomous agent engine simulates multi-turn dates citing verified facts, followed by a transparent 5-factor compatibility scoring model (interests 20%, lifestyle 25%, values 25%, communication 15%, chemistry 15%).
Next.js AI dating app. Apify scrapes ur LinkedIn/IG for a psych profile. It spawns an AI agent to go on simulated speed dates to vet matches for u. It completely automates the talking stage!
I used the apify-client SDK in a Next.js API route, running Promise.allSettled to execute concurrent scraping. I used apify/linkedin-profile-scraper for career history and apify/instagram-scraper for bio, follower metrics, and post captions. Once this raw JSON footprint is scraped, I pass it directly to OpenRouter's API, utilizing their massive-context Space Bunny Alpha model to synthesize the final psychometric profile.
Anurup R Krishnan Vercelnot aligned
The required video is missing. It was removed by the uploader, so there is no demo to grade.
Cupidly: Paste a LinkedIn or a public Instagram; an AI agent reads both, builds a needs/hobbies/interests profile, dates every other agent on your behalf, and ranks who fits you best.
Playwright uses a real browser with a logged-in session. It opens each public LinkedIn profile (headline, about, experience, posts) and each public Instagram profile (bio, captions, alt text, images) and extracts the text. Private Instagram accounts are skipped. Gemini analyzes the scraped text and images, and every claim must quote evidence found in the scraped text. Gemini also runs the agent-vs-agent.
Maddala Gayathri Rendernot aligned
Never reads LinkedIn or Instagram; zero profiles analyzed and no compatibility scores computed.
AgentMatch is an agentic dating platform where AI agents analyze public LinkedIn and Instagram signals, simulate conversations, assess compatibility, and rank potential matches.
Apify was used for public-source data extraction from LinkedIn and Instagram. The scraped profile data was processed using Node.js/JavaScript, then analyzed by the AI agent layer to extract interests, hobbies, skills, and other profile signals for compatibility matching.
Cupid Agents: paste LinkedIn+Instagram, an agent profiles each person (needs/hobbies/interests), agents date each other live, and everyone gets a ranked match list. 25 real people seeded.
Scraping: server fetch of the two public URLs with browser UA parses og:title/og:description; with APIFY_TOKEN set, route calls Apify actors instagram-profile-scraper + linkedin-profile-scraper via REST and merges posts/captions. No login, public only. Analysis/matching/dating run deterministically in lib/engine.ts (tag Jaccard + values + city, templated transcripts); optional LLM rephrase. Stack: Next.js 14 + React + API routes, deploys to Vercel.
Agentic dating site: paste LinkedIn + Instagram links, an AI agent profiles each person, agents go on simulated dates with each other, and every person gets a ranked list of best matches.
Apify actors scrape public data: harvestapi linkedin-profile-scraper for LinkedIn and apify instagram-profile-scraper for Instagram, called through the Apify REST API from a FastAPI backend. Output is trimmed of URLs and empties, then passed to Gemini, which writes a structured profile. Private Instagram accounts are rejected. Data is stored in SQLite; the front end is plain HTML/JS.
An agentic dating platform where AI agents represent 25+ real people, analyze their public LinkedIn and Instagram profiles, date each other autonomously, and generate ranked compatibility matches.
Playwright browser automation with stealth evasion: randomized user-agents, human-like scroll delays, and headless Chromium to extract LinkedIn roles, skills, bios, and posts. For public Instagram, we scrape bios, captions, and hashtags without authentication. Data feeds into our structured LLM analyzer to infer core needs, hobbies, personality traits, and dealbreakers. In production, a resilient OpenGraph/meta parser ensures 100% uptime against anti-bot rate limits.
UnDate turns public LinkedIn and Instagram profiles into AI dating agents that hold real conversations, test chemistry, explain their choices, and rank the strongest evidence-backed matches.
I used Apify actors to collect public LinkedIn and Instagram data through a Next.js backend. The raw results are validated, cleaned, and stored in Supabase before Gemini analyzes them into evidence-backed profiles. Private or unavailable profiles are never bypassed the run stops and asks the user to retry or provide profile text they are authorized to share. Groq acts as a backup if Gemini is temporarily unavailable.
Ashutosh Ranjan YouTubenot aligned
No website submitted (the live link is a YouTube URL) and the video is a music video, not a demo.
- Python, FastAPI, Pydantic, SQLAlchemy and SQLite - Next.js 14, React and TypeScript - Apify REST API for search and profile acquisition; Playwright alternative adapter for permitted public pages - Gemini 2.5 Flash for public-profile simulation; deterministic mock behavior for fixtures
maxim monastyrskiy Vercelnot aligned
No video submitted (bare youtube.com link). The site works, but the top-graded artifact is missing.
Autonomous agents date on behalf of 25 real people using only their LinkedIn and public Instagram, analyzing needs and hobbies to simulate live dates and compute fit rankings.
used a zero-auth HTTP crawler stack with User-Agent bot emulation: • LinkedIn: Googlebot/2.1 crawler headers bypass the 999 auth-wall on public profiles, extracting Schema.org JSON-LD Person data & OpenGraph tags (career, title, location). • Instagram: Facebook External Hit (facebookexternalhit/1.1) bypasses SPA login walls, parsing OpenGraph headers (og:image CDN avatar) and meta description for bio & metrics. • Pipeline: Node.js, Axios, Cheerio, and Jina AI Reader fallback.
Paste a LinkedIn + Instagram. Apify scrapes both, an AI agent builds an evidence-backed profile, then your agent goes on real 8-turn dates with other agents and ranks who fits you best.
Scraping via Apify (apify-client). LinkedIn: harvestapi/linkedin-profile-scraper (no cookies, batches of 10 on the free plan). Instagram: apify/instagram-profile-scraper (bio + latest 12 posts, captions, images). Raw JSON is stored in Supabase Postgres, and photos are rehosted to Supabase Storage. Groq runs the agents: gpt-oss-120b for analysis and verdicts, Qwen vision for photos, and gpt-oss/Qwen for the date turns. The app is Next.js 16 on Vercel.
An autonomous multi-agent dating platform where AI agents analyze real individuals' public LinkedIn & Instagram profiles to represent them on simulated dates and compute mutual compatibility rankings.
Built with Python, FastAPI, and BeautifulSoup4. Uses custom HTTP clients (httpx) with browser header spoofing and OpenGraph metadata extractors to parse public LinkedIn (headline, career, skills) and Instagram (bio, handle, visual aesthetic vibes) profiles. Extracts structured persona attributes (Needs, Hobbies, Interests, Qualities) without API rate limits, powering dual-agent simulated date dialogues and mutual compatibility ranking matrices.
Cupid Agent is an autonomous multi-agent dating site. It builds a smart "digital twin" of you using, your Instagram and LinkedIn to capture real personality, values and lifestyle for matching.
Node.js: Runs the scraping scripts for Instagram and LinkedIn Puppeteer: Opens public profile pages automatically without logging in. Cheerio: Pulls text, metadata, and work history out of the page code. HTTP / Meta Parsers: Fetches public web data, bios, and captions directly from Instagram. Using the available mode of action above made it easier to get public accessible profiles and access them easily.
Twofold turns LinkedIn and public Instagram into evidence-backed agents. They negotiate simulated dates, assess each other independently, and explain a directional ranking for every person.
TypeScript/Hono calls Apify's HarvestAPI LinkedIn Profile Scraper and Instagram Profile Scraper. Identity and public-account checks gate analysis. Extracted evidence retains source URLs and citations in SQLite/Cloudflare D1. Groq generates profile analyses, separate agent responses, and independent assessments. React and Vite power the frontend; Cloudflare Workers hosts the API and website.