Y Combinator Spring 2025 Batch: Comprehensive Market Analysis

Executive Summary

Y Combinator's inaugural Spring 2025 batch represents a strategic milestone in the accelerator's evolution, featuring 144 companies that collectively demonstrate the increasing sophistication and diversification of the startup ecosystem . This first-ever spring cohort ran from April to June and showcases Y Combinator's expansion from two to four annual batches, reflecting the accelerated pace of technological innovation, particularly in artificial intelligence .

The batch is dominated by AI-driven solutions, with 67 companies (46%) categorized as "AI agents" - representing companies building or designing tools to create autonomous AI systems . This concentration reflects broader market trends toward intelligent automation and represents an uptick from YC's Winter 2025 batch, where 58 out of 163 startups were AI agent companies .

Key themes emerging from the Spring 2025 batch include:

AI Automation Across Industries: Companies are applying agentic AI to diverse sectors from healthcare operations to manufacturing, demonstrating the technology's broad applicability
Enterprise-First Approach: A significant portion of companies target B2B markets, addressing operational inefficiencies and labor shortages in traditional industries
Regulatory-Heavy Markets: Many startups are tackling highly regulated industries like healthcare, financial services, and manufacturing, where AI can provide substantial value despite compliance complexities
Infrastructure and Developer Tools: Strong representation of companies building platforms and tools to enable other organizations to leverage AI capabilities

The diverse portfolio spans critical infrastructure sectors including space communications, autonomous manufacturing, cybersecurity, healthcare technology, and financial services, representing substantial market opportunities with companies targeting markets ranging from the $30 billion time-critical logistics sector to the expanding space economy .

Business Model Analysis

AI Automation & Agents

The largest category in the Spring 2025 batch centers on companies building autonomous AI systems to automate complex business processes. These companies typically follow SaaS subscription models or performance-based pricing tied to operational improvements.

Healthcare AI Agents: Companies like Aegis automate insurance denial appeals with an end-to-end platform that could capture value from the $260+ billion annually lost to denied claims in the US healthcare system . SynthioLabs provides AI Medical Experts for pharmaceutical companies, addressing the challenge where 7 out of 10 oncologists report no meaningful touchpoints with pharma representatives .

Industrial AI Solutions: Companies like LineWise offer AI assistants for manufacturing teams to resolve production line issues, while Mbodi AI creates embodied AI platforms allowing natural language programming of industrial robots . These solutions target the substantial market where labor accounts for nearly half of global GDP ($40+ trillion) and over 70% of factories face severe labor shortages .

Business Process Automation: Airweave enables AI agents to search any application by creating searchable knowledge bases, while Casco provides AI security testing to identify vulnerabilities in AI systems deployed across 60% of Fortune 500 companies .

Healthcare Technology Solutions

Healthcare technology companies in the batch focus on data integration, patient engagement, and provider operations, typically targeting direct-pay models or enterprise healthcare systems.

Galen AI provides personal AI healthcare agents powered by clinical and wearable data, connecting medical records from over 800 healthcare institutions across more than 12,000 locations . The platform addresses the nearly 1 billion daily health-related Google searches that receive generic responses .

Healthcare operational solutions like Aegis target provider revenue recovery through automated insurance appeals, addressing the significant gap where fewer than 15% of denials are currently appealed despite over 50% success rates .

Financial Services & Fintech

Fintech companies in the batch focus on automation and user experience improvements in traditionally complex financial processes.

Approval AI transforms homebuying by automating the entire mortgage process, positioning itself as a "Common App" for mortgages where buyers complete one form and AI handles lender interactions . Beluga Labs operates as "Rippling for Content Creators," providing comprehensive financial operating systems for the $250+ billion creator economy .

Specialized solutions like Atlog develop AI voice agents for retail stores handling in-house financing, processing 200+ sales and collection calls daily for rent-to-own furniture businesses .

Industrial & Manufacturing

Industrial companies in the batch address fundamental challenges in manufacturing efficiency, automation, and supply chain management through advanced technologies.

SAVA rebuilds sheet metal manufacturing through autonomous factories, targeting the $30 billion sheet metal industry that relies heavily on 30-year-old technology and faces significant labor shortages . The company reduces customer iteration cycles from 5-6 days to 1-2 days for clients like Pave Robotics and Zipline .

Cascade Space provides turn-key communications systems for lunar and deep space missions, offering commercial alternatives to NASA's oversubscribed Deep Space Network which is unavailable for most commercial operations .

Developer Tools & Infrastructure

Developer-focused companies provide platforms and tools to enable broader AI adoption across organizations.

Better Auth develops comprehensive authentication frameworks for TypeScript, enabling developers to implement authentication systems without relying on third-party services . Plexe builds predictive ML models from natural language descriptions, democratizing machine learning development for companies with abundant data but limited ML expertise .

Den positions itself as "Cursor for knowledge workers," creating platforms that make AI agent development accessible to non-engineers through unified interfaces combining elements of Slack and Notion .

Consumer Applications

Consumer-focused companies leverage AI to enhance personal productivity and communication experiences.

Text.ai integrates AI directly into existing messaging platforms (SMS, WhatsApp, Telegram) rather than requiring new app downloads, targeting the 5.5 billion people who use these platforms daily .

Willow develops AI dictation applications targeting professionals with heavy typing workflows, claiming 40%+ greater accuracy than built-in dictation tools with sub-second response times .

Target Market Breakdown

Enterprise/B2B

The majority of Spring 2025 companies target enterprise customers, reflecting the substantial opportunities for AI-driven operational improvements in business contexts. Enterprise-focused companies typically address markets with significant pain points, regulatory requirements, and willingness to pay premium prices for effective solutions.

Manufacturing and Industrial: Companies like SAVA, LineWise, and Mbodi AI target manufacturers facing labor shortages and operational inefficiencies. The market opportunity is substantial, with over 70% of factories experiencing severe labor shortages .

Business Operations: Solutions like Airweave, Den, and Probo focus on automating business processes, compliance management, and productivity enhancement across diverse enterprise environments.

Security and Compliance: Companies like Casco and BitPatrol address the growing enterprise need for AI security and cybersecurity solutions, particularly as 73% of enterprises experienced at least one AI-related security incident in the past 12 months at an average cost of $4.8 million .

Healthcare Providers

Healthcare-focused companies target providers, hospitals, and healthcare systems dealing with operational challenges, patient engagement, and revenue optimization.

Aegis addresses healthcare provider revenue recovery, targeting the $260+ billion annually lost to denied insurance claims . The company's platform integrates with EHRs, practice management systems, and clearinghouses to automate denial detection and appeal generation .

SynthioLabs targets pharmaceutical companies' commercial and medical teams, addressing the challenge where most physicians receive little to no engagement from pharma representatives .

Financial Services

Financial services companies in the batch target both consumers and businesses struggling with complex financial processes and inefficient service delivery.

Approval AI addresses homebuyers navigating mortgage complexity, while Beluga Labs targets the rapidly growing creator economy valued at over $250 billion . Atlog focuses on rent-to-own furniture stores that handle their own financing agreements .

Manufacturing & Industrial

Manufacturing companies face unique challenges with aging technology, labor shortages, and the need for modernization. Companies like SAVA address fundamental manufacturing processes, while Cascade Space targets the expanding commercial space industry.

The sheet metal manufacturing industry, targeted by SAVA, represents a cornerstone of American manufacturing spanning industries from Ford Trucks to NASA Space Shuttles, yet relies on 30-year-old technology .

Consumer/Direct-to-Consumer

Consumer-focused companies target individuals seeking productivity improvements, health management, and enhanced communication experiences.

Galen AI targets health-conscious consumers willing to pay for personalized AI-powered health insights, while Text.ai addresses the massive market of 5.5 billion daily messaging platform users .

Willow targets professionals including engineers at YC companies, managers at companies like GitHub and Canva, and sales teams requiring rapid communication .

Developer/Technical Audiences

Developer-focused companies target the technical community building AI applications and infrastructure.

Better Auth targets TypeScript developers seeking authentication solutions with control over their data . Plexe addresses companies with abundant data that haven't been able to build ML features due to complexity or resource constraints .

Core Technology Analysis

AI Agents & Automation

The dominant technological approach in the Spring 2025 batch involves creating autonomous AI systems capable of performing complex tasks traditionally requiring human expertise.

Natural Language Processing: Companies like Mbodi AI enable teaching robots new skills through natural language, while Plexe builds ML models from natural language problem descriptions .

Multi-Step Process Automation: Aegis automates end-to-end insurance appeals including documentation compilation and submission tracking . Approval AI handles the complete mortgage process from rate shopping to negotiation and paperwork .

Voice and Conversational AI: Atlog develops AI voice agents capable of conducting sales and payment conversations in multiple languages . Willow creates voice dictation with context-awareness for technical terms and personalization features .

Healthcare Data Integration

Healthcare technology companies focus on integrating disparate data sources to enable comprehensive patient insights and operational improvements.

Galen AI connects medical records from over 800 healthcare institutions plus over 20 popular wearable devices, leveraging state-of-the-art LLM memory capabilities for personalized health insights .

The platform integrates with Apple HealthKit and syncs with devices including Apple Watch, Fitbit, and Dexcom to provide 24/7 on-demand medical support .

Advanced Manufacturing & Robotics

Manufacturing technology companies leverage robotics, automation, and AI to address fundamental industrial challenges.

SAVA builds autonomous robotics systems specifically designed for sheet metal operations, creating robot operators that require zero programming . Mbodi AI combines generative AI, agent orchestration, and symbolic reasoning to create cloud-to-edge systems that understand environments and execute precise robot actions within seconds .

LineWise connects siloed data from PLCs, maintenance logs, and machine manuals using AI-powered diagnostic platforms to rapidly identify root causes of production issues .

Security & Compliance

Security-focused companies develop sophisticated approaches to protecting AI systems and ensuring regulatory compliance.

Casco employs AI-driven attack simulation using advanced techniques similar to expert human red teamers to devise multi-step attacks on AI systems . The platform provides human-validated expertise with evaluations supervised by teams with experience at AWS, Microsoft, and the US government .

Developer Platforms & APIs

Developer-focused companies create infrastructure and tools to democratize AI capabilities across organizations.

Airweave provides REST and MCP interfaces with Python and TypeScript SDKs, supporting multiple file formats including DOCX, PDF, PPTX, XLSX, HTML, TXT, PNG, and JPEG . Better Auth offers a framework-agnostic authentication library with plugin ecosystem and infrastructure layer .

Consumer AI Applications

Consumer applications focus on embedding AI capabilities into existing user workflows and communication patterns.

Text.ai eliminates adoption friction by integrating AI into existing messaging infrastructure, requiring no downloads or new account creation . Cactus provides comprehensive business automation including website creation, AI-powered phone answering, and booking management .

Competitive Barriers & Market Dynamics

Regulatory & Compliance Barriers

Many Spring 2025 companies operate in highly regulated industries where compliance requirements create significant barriers but also competitive moats once overcome.

Healthcare Regulations: Companies like Galen AI and Aegis must navigate HIPAA compliance and healthcare data privacy regulations. Aegis requires deep expertise in insurance processes and must maintain compliance with insurance industry standards across multiple payers .

Financial Services Compliance: Approval AI faces the challenge of building trust with both consumers and financial institutions in a heavily regulated mortgage industry, requiring regulatory compliance across different states .

Pharmaceutical Regulations: SynthioLabs must ensure all interactions are medically accurate and compliant with pharmaceutical industry marketing regulations, requiring deep expertise in both life sciences and AI .

Technical & Development Barriers

The technical complexity of building reliable AI systems creates substantial barriers that also serve as competitive advantages.

AI Model Development: Companies like Casco must maintain current knowledge of evolving AI vulnerabilities and attack methods, while Plexe faces the challenge of building automated ML pipelines that consistently produce production-ready models without human intervention .

Integration Complexity: Airweave must build and maintain reliable integrations across diverse application APIs while ensuring compatibility as third-party platforms evolve .

Robotics and Manufacturing: Mbodi AI faces the complexity of embodied AI and real-time robotics control systems, requiring deep expertise in both AI and robotics with substantial capital requirements for industrial deployments .

Market Access & Customer Acquisition

Several companies benefit from strategic partnerships and relationships that create barriers for new entrants.

Industry Partnerships: Mbodi AI has entered a joint commercialization agreement with ABB Robotics, providing market access that would be difficult for new entrants to replicate . SynthioLabs' founding team has worked with 10 of the top 20 pharmaceutical companies .

Conservative Customer Bases: Manufacturing and industrial customers are typically conservative regarding new technologies, creating adoption barriers but also customer retention benefits once integrated.

Capital & Resource Requirements

Hardware and infrastructure-intensive companies face significant capital barriers that also create competitive moats.

Manufacturing Infrastructure: SAVA faces high capital requirements for robotics development and the challenge of replacing an entire industry built on outdated technology .

Space Infrastructure: Cascade Space requires substantial capital investment for ground station network infrastructure, though the oversubscription of NASA's DSN creates significant market opportunities .

Network Effects & Platform Advantages

Several companies are positioned to benefit from network effects and platform dynamics.

Developer Ecosystems: Better Auth's open-source approach with 14k+ stars and 1.5M+ downloads creates community adoption advantages . Airweave's open-source tool enables building ecosystem of integrations and maintaining developer mindshare .

Data Network Effects: Companies like Plexe benefit from community contributions to their automated ML platforms, while Galen AI's platform improves through continuous learning about users' evolving health needs .

Market Opportunities & Outlook

Key Market Themes

AI-First Operations: The Spring 2025 batch demonstrates a clear trend toward AI-first approaches to business operations, with companies applying agentic AI across industries from healthcare to manufacturing to financial services.

Labor Shortage Solutions: Multiple companies address critical labor shortages in traditional industries, offering AI-powered alternatives to human-intensive processes in manufacturing, healthcare operations, and customer service.

Enterprise AI Adoption: The concentration of enterprise-focused companies suggests strong market demand for AI solutions that can demonstrate clear ROI and handle complex business processes.

Total Addressable Markets

The companies in the Spring 2025 batch target substantial market opportunities:

Healthcare Operations: Over $260 billion annually lost to denied insurance claims in the US healthcare system
Creator Economy: Over $250 billion market for financial services targeting content creators
Manufacturing Labor: Nearly half of global GDP ($40+ trillion) involves labor that could be augmented or replaced by AI solutions
Time-Critical Logistics: $30 billion same-day delivery market addressable by AI optimization
Global Messaging: 5.5 billion daily users of SMS, WhatsApp, and Telegram represent massive addressable market for AI integration

Strategic Implications

Vertical AI Specialization: The success of companies like Aegis, SynthioLabs, and SAVA suggests that vertical-specific AI solutions may be more successful than horizontal platforms, as they can address industry-specific compliance, workflow, and integration requirements.

Enterprise-First Growth Strategy: The predominance of B2B companies indicates that enterprise markets may offer more immediate opportunities for AI startups, given their willingness to pay premium prices for solutions that demonstrate clear operational improvements.

Platform vs. Point Solution Dynamics: Companies like Den and Airweave are building platforms to enable AI development by non-technical users, while others focus on solving specific problems. The market will likely support both approaches, with platforms capturing broader value but point solutions achieving faster initial adoption.

Regulatory Arbitrage: Companies successfully navigating regulatory requirements in healthcare, financial services, and manufacturing may build sustainable competitive advantages, as these barriers deter many potential competitors while creating substantial customer value in underserved markets.

The Y Combinator Spring 2025 batch represents a maturation of the AI startup ecosystem, moving beyond foundational AI technology toward practical applications that solve real business problems across diverse industries. The success of these companies will likely depend on their ability to demonstrate measurable ROI, navigate regulatory requirements, and build sustainable competitive advantages in their chosen markets.

References

Nearly half of the startups in [Y Combinator](https://pitchbook.com/profiles/investor/40420-36)‘s first-ever Spring cohort were companies building or designing tools to create AI agents, continuing the famed accelerator’s tendency to reflect the broader startup market’s trends. Of the 144 startups in the Spring 2025 batch, 67 are described as “AI agents” within the accelerator’s online database. That’s an uptick from YC’s last cohort, when 58 out of the 163 startups featured in its Winter 2025 batch were AI agent companies. ![chart visualization](https://public.flourish.studio/visualisation/23715197/thumbnail) Y Combinator began to deepen its focus on AI in the last couple of years as part of the boom sparked by [OpenAI](https://pitchbook.com/profiles/company/149504-14)’s introduction of ChatGPT. Of the 260 companies that participated in its Winter 2024 program, more than half were building or using AI technology.

2025 marks the first time [YC](https://www.businessinsider.com/best-startup-incubators-and-accelerators-grow-your-business-fast-2020-8) has run a spring cohort, which ran from April to June. For its inaugural spring batch, YC partners said they were especially keen on startups exploring new and creative ways to harness [artificial intelligence](https://www.businessinsider.nl/ai-has-these-founders-in-a-new-york-state-of-mind/). Many of the startups that made the cut apply agentic AI across industries — from cybersecurity to consumer fintech to healthcare. BI dug through YC's Spring 2025 cohort to find 10 [AI agent startups](https://www.businessinsider.com/generative-ai-evolution-software-companies-develop-ai-agents-workforce-2025-3) revolutionizing their respective industries. To represent the spread of startups in the overall cohort, we picked a handful of the many startups YC categorized as B2B, two healthcare startups, and one each of fintech, consumer, and industrials.

![YC AI AGENTS](https://pitchbook.brightspotcdn.com/dims4/default/852e465/2147483647/strip/true/crop/510x510+335+0/resize/1000x1000!/quality/90/?url=https%3A%2F%2Fk2-prod-pitchbook-prod.s3.us-east-1.amazonaws.com%2Fbrightspot%2F65%2Fd3%2Fd3e746ab46e7ba19fa1299191b91%2F06-11-25-ycdemoday-header-image.png) Featured image by Drew Sanders/PitchBook News Accelerators # Y Combinator is going all-in on AI agents, making up nearly 50% of latest batch ## Of the 144 startups featured in YC’s first Spring 2025 batch, 67 of them, or 46%, were AI agent companies. By [Jacob Robbins](https://pitchbook.com/news/author/jacob-robbins) & [Kia Kokalitcheva](https://pitchbook.com/news/author/kia-kokalitcheva) Published June 11, 2025 Share:

chrt is a logistics platform for time-critical shipping. We enable businesses to deliver their most critical items same-day. Time-Critical Logistics is a $30B market growing at a 20% CAGR driven by growth in American healthcare and manufacturing. Businesses that place time-critical orders are enterprises who have urgent items they need to deliver ASAP. These orders are critical in nature and often require white glove service -- think therapeutic treatments for healthcare providers or specialized equipment for manufacturers. chrt provides a marketplace of professional couriers who pick up and deliver these orders. We build the software these couriers use to handle the shipments, including a driver app and fleet management tools. This gives us full control of the shipment and allows us to provide shippers with end-to-end tracking. It also allows us to intervene early if we notice something is going wrong. We’re software engineers that became friends years ago at McKinsey.

We’re [Krishang Todi](https://linkedin.com/in/krishang-todi-a6525b135), [Aarav Bajaj](https://linkedin.com/in/aarav-bajaj-408ab01b1) , and [Dhanya Shah](https://linkedin.com/in/dhanya-shah-97878835b) and we’re building [Aegis.](https://aegishealth.us/) 💥 **TL;DR:** Aegis **automates the insurance appeals** **process end-to-end**: from denial detection and compliant appeal generation to submission, tracking, and actionable analytics. We integrate with EHRs, clearinghouses, and payer portals, helping providers and medical billing firms **recover lost revenue and save significant staff time.** [Launch Video](https://youtu.be/6AYEhYpMjBQ) **🩺 The Problem** US healthcare providers lose **$260B+** annually to denied claims, and spend **$20B+** fighting them. Yet **fewer than 15%** of denials are appealed - even though **over 50%** of appeals are successful. With AI-driven denials on the rise, the system is collapsing under its own weight.

💡 Our always-on **_AI medical expert_** delivers **_compliant, accurate, and science-first engagement_** to every physician. We help pharma GTM teams **3x their physician reach** while ensuring every interaction is medically accurate and personalized. **🎥 See us in action:** [https://youtu.be/OoJA5R0y9EQ](https://youtu.be/OoJA5R0y9EQ) **🚨 Why now?** Pharma’s current model is falling short - despite extensive sales teams, most doctors remain unengaged by pharma → **7 out of 10 oncologists** report no meaningful touchpoints with pharma → **Long-tail physicians (50–60%)** often receive little to no engagement from pharma reps → Many physicians express a **need for deeper scientific insight and relevant medical knowledge**, rather than generic promotional outreach With SynthioLabs, you can engage every physician, not just the top tier. No sales reps? No problem, we scale your team instantly. **🧠 What SynthioLabs does:** ✅ Identifies and screens physicians

**Untapped automation potential:** Hundreds of billions of dollars in manufacturing tasks, especially those with high variation, remain out of reach for automation because robots can’t adapt fast enough. This mismatch, between inflexible robots and ever-shifting production needs, leaves a massive automation gap across global manufacturing. We simply need robots to be more **versatile**. It wasn’t quite possible — **until now**. # **🧠 How Mbodi AI Helps** Mbodi is an embodied AI platform that makes robots learn like humans, with natural language. So anyone can teach robots new skills by talking to them and execute the learned skills reliably in production, in minutes. ![](https://www.ycombinator.com/media/?type=post&id=90633&key=user_uploads/1111983/cd7897e7-4fc9-4137-8c3d-2b6fe0b487bb)

**We’re building an embodied AI platform that lets anyone teach robots new skills — just by talking to them.** No code. No engineers. No week-long reprogram. Mbodi turns natural language and quick demos into precise, reliable robot actions that run in production in minutes, and adapt as often as you need. We are now rolling out real-world deployments to customers with a joint commercialization partner, [ABB](https://new.abb.com/products/robotics), the world’s leading robot manufacturer. [https://youtu.be/CwRxP3\_F9MA](https://youtu.be/CwRxP3_F9MA) # 🚨 **The Problem** Manufacturing is facing a critical inflection point. **Labor crisis:** Labor accounts for nearly half of global GDP ($40+ trillion), but over 70% of factories face severe labor shortages. **Flexibility gap:** Traditional robots require week-long reprogramming by engineers or specialists every time things change.

Ian Saultz ### Ian Saultz, Founder & CTO Founder & CTO of Casco [Casco](https://www.ycombinator.com/companies/casco) Founded:2025 Batch:Spring 2025 Team Size:2 Status: Active Primary Partner:[Dalton Caldwell](https://www.ycombinator.com/people/dalton-caldwell) Company Launches Casco - Agentic Red Teaming for AI agents and apps [See original launch post](https://www.ycombinator.com/launches/NZo-casco-agentic-red-teaming-for-ai-agents-and-apps) Hi everyone - we’re Rene and Ian, ex-AWS AI & Security, and we’re thrilled to introduce [Casco](https://casco.com/)! Casco prevents AI agents and apps from messing up by evaluating them for security, safety, and accuracy. Casco already secures AI systems deployed in 60% of Fortune 500 companies. [https://youtu.be/SWPnB-GADdE](https://youtu.be/SWPnB-GADdE) **The Problem**

Galen is an AI-powered personal healthcare companion. We enable patients to securely connect their medical records from over 800 healthcare institutions across more than 12,000 locations in the United States, United Kingdom, and Canada. We also connect with over 20 popular wearable devices including Apple Watch, Fitbit, and Dexcom, bringing all your medical records and health data into one unified space. We founded Galen because for over four years, Viraj struggled to navigate several conditions related to chronic low back pain. He often found himself waiting weeks for responses from his care team, received little guidance or education about new diagnoses, and was overwhelmed trying to manage overlapping conditions and medications. His experiences are not unique. Our healthcare system is overloaded and fragmented, and often, patients are left to navigate their care on their own.

Everyday there are nearly 1 billion health-related Google searches, but the responses to such queries are generic and impersonalized, because the searches lack any insight into a users personal health background. By leveraging your own data and state-of-the art LLM memory capabilities, Galen delivers personalized insights, proactive health recommendations, and 24/7 on-demand medical support — empowering patients to be an active part of their health team. We are a consumer app in beta testing and will be available on the App Store soon!

Approval AI reimagines how homebuyers get a mortgage. With one simple form, our AI assistant handles the entire process, so buyers get the best loan with less work. [https://youtu.be/PCUBgSjUvNw?si=qXIssoIa\_zCaaWH3](https://youtu.be/PCUBgSjUvNw?si=qXIssoIa_zCaaWH3) **📄 One form. Every lender.** Like a “Common App” for mortgages, buyers complete one form and our AI handles the rest. It finds the best lenders for a buyer’s profile, shares the necessary details, and guides them seamlessly from pre-approval to closing. 📞 **AI-powered negotiation** Our AI assistant reaches out to banks, brokers, and credit unions to request quotes, negotiate rates, and compare offers, getting buyers the best deal with zero awkward calls. 🤖 \*\*Smart Guidance, Human Support.\*\* The AI assistant is available 24/7 to answer questions and explain mortgage terms in plain language. When more help is needed, human experts are just a message away.

The rapidly growing creator economy ($250B+) consists of millions of solo entrepreneurs still managing complex business finances with spreadsheets and a patchwork of disconnected tools. They face volatile multi-source income, unique business expenses, and daunting U.S. tax obligations: leading to significant stress, lost time, and missed financial opportunities. Beluga Labs is building the financial operating system for these modern solopreneurs. Our platform starts with intelligent bookkeeping: we connect all their financial accounts, use AI to automatically categorize income and expenses, and assess their potential tax deductibility, giving creators a clear picture of their business finances and tax preparedness. This core solution, already actively used for expense management and tax Q&A, is the first step.

[Atlog](https://www.ycombinator.com/companies/atlog) Founded:2024 Batch:Spring 2025 Team Size:3 Status: Active Primary Partner:[David Lieb](https://www.ycombinator.com/people/david-lieb) Company Launches Atlog: AI Employee for Furniture Stores [See original launch post](https://www.ycombinator.com/launches/NbU-atlog-ai-employee-for-furniture-stores) **The Problem**: Furniture stores, specifically rent-to-own furniture stores, write their own contracts regarding rental agreements. This gives them a unique advantage: they can rent to anyone, regardless of credit score. However, this market has one major flaw: **Customers are always late with payments.** The stores know about this, and it is just a fact of the business model, but these constant late payments cause a few key operational issues:

SAVA is rebuilding sheet metal manufacturing from the ground up. From Ford Trucks to NASA Space Shuttles, sheet metal is the cornerstone of American Manufacturing - but the industry runs on 30-year-old technology and a shrinking workforce. Alessio and Vedic competed in high school robotics when they met Jakob at Georgia Tech's Robotics labs. After witnessing Vedic's father's 12-hour shifts and visiting manufacturers across the US, they saw an industry trapped by legacy constraints: manual processes, skilled labor shortages, and unused automation. SAVA is building a fully integrated manufacturing system that eliminates skilled operators while delivering superior speed, quality, and economics. They're already reducing customer iteration cycles from 5-6 days to 1-2 days for companies like Pave Robotics and Zipline - proving the technology works in real production environments.

The Deep Space Network (DSN), NASA’s global network of ground terminals for communicating with deep space probes, is [oversubscribed by 40% and in desperate need of upgrades.](https://newspaceeconomy.ca/2024/09/02/nasas-deep-space-network-challenges-and-the-road-ahead/) The DSN is all but unavailable for most commercial operations. Companies with missions going to the Moon and beyond have to deal with a patchwork of individual dish operators for every mission, each with bespoke hardware, specs, and interfaces. Furthermore, many of these commercial providers are not able to handle the demand and [operational failures also can lead to loss of mission](https://www.astroforge.com/updates-collection/odint-mission-debrief). ![](https://www.ycombinator.com/media/?type=post&id=90423&key=user_uploads/222992/c6aa8968-e567-42e5-8866-866414630f22) **Comms system design is fraught with peril**

### The authentication framework for TypeScript Better Auth is a compressive authentication framework for TypeScript that lets you implement everything from simple auth flows to enterprise-grade systems directly on your own database, embedded in your backend. It’s loved by developers all over the world and endorsed by leading voices in the ecosystem. On top of the framework, Better Auth also provides an infrastructure layer to help you scale with user management and analytics, bot and fraud detection, transactional email and SMS, global session storage and more. Active Founders

Lots of great use cases for ML models in businesses never materialize because ML projects are messy and convoluted. You spend months finding the data, cleaning it, experimenting with models and deploying them to production, only to find out that the project has been binned due to taking so long. At a previous employer, we witnessed a team of 10 ML engineers spend 2 years and $3M building models for a project that never saw the light of day. There are several tools for “automating” ML, but it still takes teams of ML experts to actually productionize something of value. And we can’t keep throwing LLMs at every ML problem. Why use a generic 10B parameter language model, if a logistic regression trained on your data could do the job better? ## **✅ 2. Solution** Step 1: Connect your data sources Step 2: Describe your problem statement Step 3: Use the deployed model via API [https://youtu.be/zByvH0wTSuE](https://youtu.be/zByvH0wTSuE) ## **🚀 3. How does it work?**

[Den](https://www.ycombinator.com/companies/den) Founded:2025 Batch:Spring 2025 Team Size:2 Status: Active Location:San Francisco Primary Partner:[Tom Blomfield](https://www.ycombinator.com/people/tom-blomfield) [![X (Twitter) logo](https://www.ycombinator.com/images/social/x-logo.svg)](https://x.com/supportingden) Company Launches Den: Cursor for knowledge workers [See original launch post](https://www.ycombinator.com/launches/NZ9-den-cursor-for-knowledge-workers) ### **TL;dr** Den is Cursor for knowledge workers. We make it 10x easier to build AI agents within a Slack/Notion-like platform. [https://youtu.be/OBpZ4RaV\_9Y](https://youtu.be/OBpZ4RaV_9Y) ### **The Problem** AI agents are great, but impossible for most people to use. At our previous startup, engineers would ship features during their lunch break using Cursor, while the rest of the team drowned under repetitive tasks. We believe everyone, not just engineers, deserves 10x leverage on their time.

[Jobs](https://www.ycombinator.com/companies/text-ai/jobs)0 [https://www.text.ai/](https://www.text.ai/) * * * ### AI in your group chats You can now text AI in your existing chats. No downloads needed. We've brought AI directly into SMS, WhatsApp, and Telegram which over 5.5 billion people use daily. No learning curve, just text and go. Discover local eats. Plan group events. Roast friends. Together with AI. Active Founders

What makes Willow special is that it’s not just basic voice typing. It’s smart, fast, and sounds like _you_. Here’s how: * **Context-awareness**: Willow looks at what you’re working on so it can get technical terms, names, and phrases right. * **Custom dictionaries**: You can add your own words—like company names, industry terms, or team slang—so they’re always spelled correctly. * **Smart correction and formatting**: Willow automatically fixes mistakes and adds structure—like formatting emails, breaking up paragraphs, and handling punctuation. * **Personalization** _(beta)_: We’re working on making Willow sound just like you—adjusting tone and style depending on what you’re writing. * **Privacy**: Willow runs in the cloud for speed and accuracy, but we never store or collect your voice or data. Privacy is our top priority, and it’s built into everything we do. There’s an optional, opt-in setting for users who want to help improve the model, completely anonymized.

Casco helps companies confidently ship AI agents and apps by simulating attacks on their AI systems, identifying vulnerabilities, and providing guided remediation. Companies building on AI face a disproportionate amount of cybersecurity threats. 73% of enterprises experienced at least one AI-related security incident in the past 12 months, with the average incident cost amounting to $4.8m. It also takes 40% longer for these companies to identify and address these breaches vs. non-AI related incidents. Check if your AI system is secure at casco.com --- We already secure AI systems deployed in 60% of Fortune 500 companies and are default alive. \[1\] \[1\] http://www.paulgraham.com/aord.html Active Founders

* **Detect denials automatically:** Aegis continuously monitors claims data streams from your clearinghouse and instantly flags denied claims without requiring manual intervention. These denied claims are then arranged by priority based on deadline and claim value. * **Pull patient records, EOBs, and payor docs:** Aegis integrates with the EHR to fetch supporting materials, such as patient records, Explanation of Benefits (EOBs), and payer correspondence. * **Generate and submit appeal packages:** Using AI trained on successful appeal language and payer requirements, Aegis drafts customized appeal letters, compiles required documentation, and submits the appeal directly to the insurance provider via appropriate channels (fax, portal, or clearinghouse). * **Track appeal progress and surface denial insights:** Aegis provides end-to-end tracking of every appeal, including submission confirmation, status updates, and final resolution.

From Ford Trucks to NASA Space Shuttles, sheet metal is the cornerstone of American Manufacturing. We are addressing the critical labor shortage in US sheet metal. _Our ask is two-fold:_ 1. [Order custom metal parts](https://savarobotics.com/yc.html) produced autonomously by SAVA. 2. Introduce us to family and friends in domestic manufacturing. * * * ## Footer [Y Combinator](https://www.ycombinator.com/ "Y Combinator") ### Programs * [YC Program](https://www.ycombinator.com/about) * [Startup School](https://startupschool.org/) * [Work at a Startup](https://www.ycombinator.com/jobs) * [Co-Founder Matching](https://www.ycombinator.com/cofounder-matching) ### Company

It’s 40%+ more accurate than built-in dictation tools, works in under a second, and is so good that most users replace nearly all their typing with it. ### **Who’s using Willow** People with heavy typing workflows love Willow because it helps them move faster and get more done. This includes: * **Engineers at YC** companies use Willow to prompt AI tools like Cursor up to 4x faster. * **Managers at Github, Canva, and Webflow** rely on it to brainstorm, take notes, and write docs hands-free. * **Sales teams** use Willow to reply instantly on Slack, email, and CRM tools. * **Busy professionals** drowning in email use Willow to clear out their inbox in minutes, not hours. ### **Our ASK** **Try Willow Voice** → [WillowVoice.com](http://willowvoice.com/) It helps you get work done 4x faster, just by talking rather than typing. It’s free to start and works instantly on Mac. We’d love to hear what you think!

[Better Auth](https://www.ycombinator.com/companies/better-auth) Founded:2025 Batch:Spring 2025 Team Size:2 Status: Active Location:San Francisco Primary Partner:[Pete Koomen](https://www.ycombinator.com/people/pete-koomen) [![X (Twitter) logo](https://www.ycombinator.com/images/social/x-logo.svg)](https://twitter.com/better_auth) Company Launches Better Auth - The Authentication Framework for TypeScript [See original launch post](https://www.ycombinator.com/launches/NUm-better-auth-the-authentication-framework-for-typescript) Hey! 👋 We’re building Better Auth - a comprehensive, framework-agnostic authentication framework for TypeScript. It lets you roll your own auth directly on your backend and database, without relying on 3rd party services. From basic sign-ins to complex enterprise systems all with great DX. [https://www.youtube.com/watch?v=j\_HjHmprVOs](https://www.youtube.com/watch?v=j_HjHmprVOs)

## **🙏 5. Our ask** * We’d love to get in touch with any companies that have a lot of data but have not been able to build ML features on top of them! * If you’re an individual who believes that the future of ML development is autonomous and open source, check out our agent: [https://github.com/plexe-ai/plexe](https://github.com/plexe-ai/plexe) Get in touch with us: [founders@plexe.ai](mailto:founders@plexe.ai) ## Footer [Y Combinator](https://www.ycombinator.com/ "Y Combinator") ### Programs * [YC Program](https://www.ycombinator.com/about) * [Startup School](https://startupschool.org/) * [Work at a Startup](https://www.ycombinator.com/jobs) * [Co-Founder Matching](https://www.ycombinator.com/cofounder-matching) ### Company

Teach robots new skills through natural language and run reliably in production in minutes. Mbodi combines generative AI, agent orchestration, and symbolic reasoning, turning everyday language and quick demos into precise, reliable robot actions for industrial automation. Active Founders

[https://plexe.ai](https://plexe.ai/) * * * ### Open-source agents to build predictive ML models from a prompt Plexe builds predictive ML models from a problem description. It connects to data sources, conducts experiments, evaluates and deploys the models to an API endpoint. Active Founders

* The agent speaks in **English** and **Spanish** * **Optimizes call time** and day to improve conversion * **Finds updated contact info** (customers often change numbers or addresses) * **Collects payments** directly over the phone **Our background:** Shaun and Vraj both grew up in communities where rent-to-own furniture was a household name, and our families even relied on it when they immigrated to the U.S. We wanted to build for people we actually enjoy working with. The furniture rental industry is one of those rare industries where we get to travel, solve real problems, and meet people we relate to. **Our Ask**: If you know anyone in the furniture industry or are tangentially related to anyone in the furniture industry, we’d love to talk to them! Our focus is specifically on rent-to-own furniture, but if they work in other aspects of the business, that is ok as well! ## Footer [Y Combinator](https://www.ycombinator.com/ "Y Combinator") ### Programs

Meet Galen AI — our AI-powered solution to make high-quality, personal medical support accessible to everyone. No more screenshotting your records and uploading them to ChatGPT. No more WebMD rabbit holes about your symptoms. No more waiting months for your next doctor’s appointment. ![](https://www.ycombinator.com/media/?type=post&id=90700&key=user_uploads/1833511/b89751e0-80e4-479d-aa23-bf7209579b13) **All Your Medical Data in One Place** Galen securely integrates with **Apple HealthKit** to connect directly to your medical records, allowing you to unify health data from all your providers and institutions in one place. It also syncs with your **Apple Watch** to bring in daily fitness and activity data, with support for **20+ additional wearables** coming soon. As you share new symptoms and health updates, Galen **keeps track** and **continuously learns** about your evolving needs, providing long-term support.

[Spring 2025](https://www.ycombinator.com/companies?batch=Spring%202025)[Industrials](https://www.ycombinator.com/companies?industry=Industrials)[Manufacturing and Robotics](https://www.ycombinator.com/companies?industry=Manufacturing%20and%20Robotics) [![](https://bookface-images.s3.amazonaws.com/small_logos/fe1589548b774d33ca4f20c5020467a9627f9c13.png)](https://www.ycombinator.com/companies/godela) [GodelaSan Francisco, CA, USA AI Physics Engine to replace simulations and prototypes](https://www.ycombinator.com/companies/godela) [Spring 2025](https://www.ycombinator.com/companies?batch=Spring%202025)[B2B](https://www.ycombinator.com/companies?industry=B2B)[Engineering, Product and Design](https://www.ycombinator.com/companies?industry=Engineering,%20Product%20and%20Design) [![](https://bookface-images.s3.amazonaws.com/small_logos/a49803acbbc34d447ce47d0129bde3f02dd6b0ec.png)](https://www.ycombinator.com/companies/bitpatrol) [BitPatrolNew York, NY, USA

We’ve built a cloud-to-edge system that takes your instructions, understands the environment, makes a plan, and turns it into precise, safe robot actions. All within a second. Our universal agent framework can run across any robotic hardware, coordinating perception, reasoning, planning, and control in real time. * 🕐 No week-long reprograms * 👷‍♂️ No on-site engineers * 🗣️ Just say what you want done * 🏭 Run reliably in production When one robot learns something, the entire fleet does. Individual skills can be reused, repurposed, and adapted instantly — across machines, factories, and tasks. # 🚀 **What’s Happening** We’re moving toward real-world deployments after entering a joint commercialization agreement with ABB Robotics, one of the world’s leading robot manufacturers, following our win in their global AI Startup Challenge from among hundreds. This represents a first-of-its-kind system in industrial automation.

When a line goes down, it can take **hours** to identify the problem. Engineers and technicians chase PLC logs, flip through outdated manuals, and rely on the one person who “just knows.” While production stalls, tribal knowledge disappears, and the same issues repeat. ### **THE SOLUTION** LineWise centralizes critical knowledge by integrating siloed data from IoT sensors, maintenance records, and machine manuals. It ranks the most likely causes and guides the team to a fix in just **minutes**. But the best part? It remembers every issue for next time and advocates for preventative actions. * * * ### **How LineWise Works** 🔧 Upload an error code image or describe the issue in a message. 🤖 The AI provides a refocused problem statement for confirmation. 📊 The AI dynamically ranks possible hypotheses of causes by probability. 🧠 A human validates each hypothesis to identify the true cause. 🛠️ Resolve the issue by following a step-by-step guide.

1. **Sophisticated, AI-Driven Attack Simulation:** Casco uses advanced techniques, just like an expert human red teamer, to devise and execute multi-step attacks against your AI agents and apps. 2. **Human-Validated Expertise:** The AI evaluations are supervised and validated by our team, ensuring high-fidelity findings. We’ve built and secured AI systems, applications, and cloud infrastructure at AWS, Microsoft, and the US government. 3. **Actionable, Compliance-Ready Reports:** Casco provides comprehensive reports with reproduction steps, risk explanations, and remediation guidance, aiding compliance with SOC 2, NIST AI RMF, EU AI Act, and ISO 27001. **Our Story** As founding members of the Generative AI team at AWS (Lead Engineer + Security, Head of Product), we witnessed the evolving landscape of AI agents and applications. We identified emerging security gaps as traditional security practices struggled to keep pace with this new paradigm.

* **Open-Source** * **Simple yet powerful search interface** via REST or MCP * **MCP-compatible:** Turn any app into a semantically searchable MCP server * [**Python & TypeScript SDKs**](https://docs.airweave.ai/sdks) * **Connect to your agent to your favorite apps** like Google Drive, Gmail, Outlook, Google Calendar, HubSpot, Notion, Jira, Slack, Dropbox, Stripe, Github, etc. * [**White-Labeled Multi-Tenancy**](https://docs.airweave.ai/white-labeling) to provide OAuth2-based integration across user accounts with secure boundaries * **File Handling:** Supported formats: DOCX, PDF, PPTX, XLSX, HTML, TXT, PNG, JPEG # **Examples of How Customers Use Airweave:**

To take it further, and after seeing the pain points of our current users, we're building an **infrastructure layer** on top of the framework to extend the library with functionality we can't offer through the framework alone: * A comprehensive admin dashboard with user management & analytics * Bot/fraud/abuse detection * Global session storage * and more… To be clear – we’re not building a third-party auth service. Our goal is to make rolling your own auth so ridiculously easy that you’ll never need one. Get started with the library here: [better-auth.com/docs/installation](http://better-auth.com/docs/installation) Join the waitlist for infra: [http://better-auth.build/](http://better-auth.build/) Star the repo: [https://github.com/better-auth/better-auth](https://github.com/better-auth/better-auth) ## Footer [Y Combinator](https://www.ycombinator.com/ "Y Combinator") ### Programs

[https://www.youtube.com/watch?v=SFOBlc6extw](https://www.youtube.com/watch?v=SFOBlc6extw) We also just made the top 5 on [Product Hunt](https://www.producthunt.com/posts/text-ai) this past week! # **✅ How to try it:** 1\. Save our contact card: \[[text.ai](http://text.ai/)\] 2\. Text directly or add to your group chats 3\. That's it! Works on \[[SMS](https://www.text.ai/)\], \[[WhatsApp](https://wa.me/15626840069?text=What%20is%20the%20best%20Hawaiian%20city%20to%20visit%20for%20Spring%20Break%3F)\], | \[[Telegram](https://t.me/thetextai_bot)\] ![](https://www.ycombinator.com/media/?type=post&id=89968&key=user_uploads/310607/11b0f146-9cef-4d39-9e68-63318cc9299f) **🤔 Why did we build this?** Ever tried explaining to your mom how to download and use a new AI app? Yeah... that's why we built [text.ai](http://text.ai/).

Cactus ([https://oncactus.com](https://oncactus.com/)) answers the phone, qualifies leads, and handles bookings so the business owner has a full schedule of work. It is the AI Business Assistant that manages the admin so the owner can focus on what they love! Cactus has empowered 100+ chefs and is growing rapidly! Next, we’ll expand to photographers and personal trainers. **Check out our [launch video](https://youtu.be/VWeQaySQnY8)** **1️⃣ We provide every solopreneur with a website** ![](https://www.ycombinator.com/media/?type=post&id=90199&key=user_uploads/423660/d1d7e29b-8f13-4677-aa86-4381df359716) **2️⃣ Our AI answers the phone and qualifies customers.** ![](https://www.ycombinator.com/media/?type=post&id=90199&key=user_uploads/423660/3cdfc4f0-2b3e-4692-8126-247d4e5650e8) **3️⃣ And help solopreneurs manage bookings** ![](https://www.ycombinator.com/media/?type=post&id=90199&key=user_uploads/423660/af86c212-6e6d-428d-ae06-bce3cba64763) # **🎯 The Team**

Plexe uses a self-correcting team of ML engineering agents to research, experiment, evaluate, refine, and deploy the best performing model over an API endpoint. It connects to data sources, discovering relevant fields, and autonomously building the model. Think of it like your own ML helper helping you go from idea to deployed models 10x faster. [https://youtu.be/bUwCSglhcXY](https://youtu.be/bUwCSglhcXY) ## **👋 4. Team** [Vaibhav Dubey](https://bookface.ycombinator.com/user/2260416) (ex-Proofpoint, Expedia, Imperial College London) and [Marcello De Bernardi](https://bookface.ycombinator.com/user/2337737) (ex-AWS, Expedia, Oxford) met 6 years ago while building a chatbot that served 9+ million customers of a large bank. Since then, they’ve built enterprise-scale ML solutions that serve billions of predictions per day for millions of users worldwide. ![](https://www.ycombinator.com/media/?type=post&id=90589&key=user_uploads/2260416/4e5a0513-e013-4adc-9520-ebde07c44981)

**Supreet Deshpande**, **Sahitya Sridhar** and **Rajashekar Vasantha**, co-founders of **SynthioLabs**, are building voice AI medical reps for pharma commercial and medical teams to engage with their most important stakeholders—doctors and patients. Pharma spends billions annually on sales and medical representatives, yet most physicians—especially those in community settings, who make up the majority of the market—never hear from one. The result? Missed engagement, slower adoption, and lost revenue where it matters most. SynthioLabs is building voice AI medical reps that help pharma scale physician engagement in ways traditional field teams cannot. Their AI reps deliver compliant, on-demand medical info —and capture rich data on physician needs and concerns at scale. The team includes ex-McKinsey and ex-ZS life sciences experts, as well as AI researchers from AWS. They have worked with 10 of the top 20 pharma companies and bring over a decade of experience in life sciences.

They're not just replacing human operators with robots - they're building the foundation to replace an entire industry built on outdated technology with modern manufacturing intelligence.

By transforming how pharma connects with every physician—not just the top tier—SynthioLabs is set to revolutionize medical education, compliance, and outreach at scale. * [Text.ai](https://www.text.ai/) brings AI directly into SMS, WhatsApp, and Telegram, letting over 5.5 billion people access powerful AI without downloads or a learning curve. By embedding AI where people already chat, text.ai is making advanced technology instantly accessible and collaborative for everyone, everywhere. * [Better Auth](https://www.better-auth.com/) is the fastest growing auth library with 14k+ stars and 1.5M+ downloads. They let developers build secure, enterprise-grade authentication in TypeScript without relying on third-party services. This empowers teams to innovate faster while maintaining full control over their user data.