Ruby
Golang
.Net
MCP
C++
At API World 2026, we wanted developers to do more than just hear about SerpApi. So we put our APIs in their hands and let them build.
At our booth, developers could experiment with SerpApi live, connect search data to their existing tools and workflows, and see what they could create with real-time web data. Many developers participated in our hackathon and used SerpApi to build real-time search capabilities, automated workflows, and AI-driven applications.
The result? A collection of 124 amazing projects that showed just how many ways developers can put search data to work.
Check out five of the many interesting projects built with SerpApi at API World.
1. AccessForm
Call. Talk. Your form is filled.
The people who most need help with official paperwork are the ones least able to do it online. A blind person with a hospital bill, a 70-year-old who can't walk to the bus, a student with ADHD facing a disability office form. They all have a phone. Most of them don't have a screen they can use, and some don't have internet at all.
How AccessForm helps
You call a number and say what's going on in your own words. AccessForm works out which official program fits, finds the real application form from the official source, asks you its questions one at a time in plain language, fills the actual PDF, and texts you the result with a list of what's still missing.
Everything happens in one phone call. It never submits, never signs, and never decides eligibility. If it can't verify an official form for the exact organization you named, it says so and stops.
How SerpApi was used
Vapi runs the voice agent on a Twilio number. Every tool call hits a Next.js backend that uses SerpApi to find and verify the official form, OpenAI to read the form's fields and turn them into spoken questions, pdf-lib to fill the real PDF, Nutrient for accessibility tagging when credits allow, and Xano as the system of record for cases, answers, and completeness. A conversation page rebuilds the call live from Xano's event stream, so a laptop can watch a phone call as it happens.
The developers used SerpApi to search for and discover official application forms (PDFs) based on a user's specified need, organization, and location (e.g., executing live web queries to find official .gov, .edu, or official organizational site links).
Built with: SerpApi's Google Search API
👉 See project details
👉 Github
2. PantryProof
PantryProof turns fragmented recall evidence into cited, human-reviewed actions and a replayable closure packet for every partner pantry.
Food banks and partner pantries form the “true last mile” of a food recall. Products may arrive through central distribution, retail rescue, local purchasing, food drives, and other paths. Partner agencies may operate intermittently or rely on volunteers, so sending an alert does not prove that the affected product was found, contained, or communicated downstream.
PantryProof was built around a simple question: How can a food bank prove that every partner completed the correct recall response without letting an uncertain AI prediction make the final safety decision?
How PantryProof helps
PantryProof streamlines food safety responses across partner pantries by discovering official recall notices via SerpApi, extracting cited data, requiring human approval for extracted fields, and deterministically matching identifiers (UPC/GTIN and lot numbers) against inventory to trigger necessary quarantine, disposal, or notification actions before generating a verified, audit-ready PDF closure packet.
The browser experience is built with TypeScript, JavaScript, semantic HTML, and responsive CSS. A deterministic TypeScript domain engine owns identifier normalization, matching policy, agency aggregation, human review state, response completeness, and closure readiness.
How SerpApi was used
Developers used SerpApi to get structured, current Google News recall results ordered newest first. They limit results to FDA and USDA domains. PantryProof validates every returned URL, determines evidence tier from the actual hostname rather than an untrusted label, and allows a coordinator to attach a reviewed result as canonical evidence.
This makes web search part of a traceable evidence workflow instead of treating the first search result as truth.
Built with: SerpApi's Google News API
👉 See project details
👉 Github
👉 See it live
3. ClinicalBrief AI (by LearnHive Labs)
Upload patient lab data - GPT-4o searches PubMed, retrieves clinical guidelines, and delivers a signed clinical brief to the clinician in under 60 seconds.
Clinicians spend 2+ hours per complex patient case searching literature, cross-referencing lab values against treatment protocols, and writing structured briefs. Most existing AI tools either hallucinate clinical recommendations or cost $50,000+ per year. There is no open, inspectable, evidence-grounded pipeline that any clinic can run.
How ClinicalBrief AI helps
ClinicalBrief AI was built based on a research publication on adaptive health orchestration to prove that a fully agentic, source-cited, confidence-gated clinical intelligence pipeline can be built on open infrastructure - in n8n, without a single line of server code.
A patient or pharmacy staff member fills a Google Sheet with lab values: HbA1c, LDL, BMI, blood pressure, medications, diagnoses. No app. No login. No technical knowledge required.
How SerpApi was used
When a new row is added, the n8n pipeline fires automatically - SerpApi queries Google Scholar in real-time using the patient's diagnoses. Three current PubMed publications are retrieved and injected into the agent prompt before any generation happens. GPT-4o generates clinical briefs using patient data, PubMed evidence, and clinical guidelines. Briefs with a confidence score of 80% or higher are electronically signed and delivered to clinicians, while those below are reviewed by humans. All data is sent to a physician.
Built with: SerpApi's Google Scholar API
👉 See project details
👉 Github
4. ApplyPilot
ApplyPilot finds real opportunities and proves why they match you - using grounded evidence, deterministic scoring, bounded AI, and explicit uncertainty instead of black-box recommendations.
Most AI job-matching systems give candidates a score. But a score can hide a dangerous assumption: when evidence is missing, an AI system may treat that missing information as if it were negative evidence. That idea became the foundation for an evidence-controlled opportunity intelligence system where AI can interpret evidence, but cannot silently manufacture it.
How ApplyPilot helps
ApplyPilot uses a multi-step process to match candidates with real job opportunities. It extracts evidence from résumés, verifies it, and then searches the web for relevant jobs. The final result is a ranked list of opportunities with detailed explanations of why they are a good fit.
The end-to-end workflow is:
Résumé → grounded evidence → human review → candidate profile → live web discovery → deterministic evaluation → bounded AI reasoning → self-audit → ranked opportunities → Why this job?
How SerpApi was used
ApplyPilot uses SerpApi's Google Jobs API to discover real public opportunities during a Mission. The system preserves the provider query, raw listing evidence, canonical opportunity, and source lineage. Repostings can be deterministically deduplicated while preserving the underlying sources.
SerpApi helps ApplyPilot operate on changing real-world opportunities rather than a static demonstration dataset. Live web data changes the AI experience directly: the reasoning pipeline evaluates opportunities discovered during the run instead of answering from stale or fabricated listings.
Built with: SerpApi's Google Jobs API
👉 See project details
👉 Github
👉 See it live
5. DepositCheck
DepositCheck protects renters from listing scams by using SerpApi's Google Lens API to perform visual reverse image searches on listing photos, matching them against the live web to confirm if an apartment matches its claimed address or belongs to another property.
Rental scams work by theft, not invention. A scammer copies the photos and the description from a real listing, swaps the contact details and the address, and reposts it somewhere with weaker verification. Someone wires a deposit for a property the "landlord" has never owned.
The photos are the one thing the scammer cannot change. Swap them out and the listing stops looking like the flat that attracts victims in the first place. But a renter cannot check that. You would have to recognize one apartment interior among billions of published images, from memory, on a phone, while a "landlord" tells you three other people are viewing tonight.
How DepositCheck helps
You upload one photo from the listing and type the address you were given. SerpApi's google_lens engine with type=exact_matches runs Google's visual matching model over the live web index and returns every page carrying that same image. The claimed address is matched against what those pages say. The photo is deleted as soon as the lookup returns.
One of three verdicts is generated - CORROBORATED (a genuine syndicated listing), CONTRADICTED (the photos belong to one specific other address), or UNVERIFIED (not enough evidence either way).
How SerpApi was used
DepositCheck uses SerpApi's Google Lens API with type=exact_matches to return every page carrying that same image. The claimed address is matched against what those pages say. The team aims to extend the solution in the future by using Google Maps API to verify the address is a residential property rather than a parking lot or a mailbox storefront and SerpApi's Google Search API on the landlord's phone number and email, which scam operations reuse across dozens of listings.
Built with: SerpApi's Google Lens API, Google Maps API (Future Extension), and Google Search API (Future Extension)
👉 See project details
👉 Github
👉 See it live
What Developers Are Building With Search Data
The projects built at API World were very different from one another, but they had something in common:
Developers were using search as a building block for something bigger
Some projects focused on AI and agents, using search results to give applications access to current information. Others used search data for automation, research, data collection, and discovery.
And because SerpApi returns structured search results in JSON and Markdown, developers can take that data and plug it directly into the applications and workflows they're already building.
Check out all the projects developers built with SerpApi at API World here:
The projects built at API World are just a small sample of what's possible.
Build with SerpApi
For us, the most exciting part wasn't simply seeing developers call an API and build a tool using it. It was seeing what they built once they had access to real search data.
Whether you're building an AI agent, research tool, automation, data pipeline, or something completely different, live search results can be an important source of fresh information for your application.
And you don't have to build the search infrastructure yourself.
With SerpApi, developers can access Google and other search engines through an easy and fast API and work with the results in a format that's ready to use.
Have an idea? Start building.
Building something useful with SerpApi? Share it with us.
Interested in exploring our use cases? Check out our use cases page.
Read Some Of Our Latest Posts
Adarsh Divakaran
Musa Tunçkılıçoğlu
Noraina Nordin
