Computer Vision Startups funded by Y Combinator (YC) 2025

Matterport is an immersive media technology company that is shaking up the 3D / VR world. Our team has built the first end-to-end system for creating, modifying, distributing, and navigating immersive 3D and virtual reality (VR) versions of real-world spaces on web and mobile devices. Matterport offers the world's most inexpensive and simplest way to capture 3D spaces. Our products include: - Matterport Pro Camera for capturing real spaces in 3D. It collects accurate visual and spatial data to map entire areas in minutes and is all about automation and ease of use. - The Matterport Cloud for processing and hosting 3D models - Matterport Portal, our system for viewing, editing, and managing models; collaborating with colleagues; and sharing models with others - Matterport 3D Showcase, a browser-based 3D media player, which allows anyone to view 3D models in their browser with no additional software - Matterport Core VR: All Spaces can be converted to VR and experienced on Samsung Gear VR or Google Cardboard (in beta), with additional device support coming soon. Matterport 3D media solutions power industries from real estate (residential, multi-family and commercial) and travel and hospitality (hotels, vacation rentals, and venue booking), to business listings, architecture, engineering and construction, news and entertainment, and everything in between. We’re growing fast. If you’re passionate about solving cutting-edge problems in computer vision and hardware design and creating order-of-magnitude improvements in the ability to easily create and share 3D models of real world spaces, we want to talk to you. See open positions at matterport.com/jobs. Try Matterport for yourself at matterport.com/try.

Efference builds the eyes and visual cortex for robots. Instead of treating depth as a hardware problem, we use software — inspired by how humans see — to generate rich and reliable 3D information. This lets us deliver higher performance and greater ease of use at a lower price than existing stereo cameras.

With our visual understanding engine & platform, Aspect primarily serves both enterprise content teams and data teams at AI labs. For enterprise: Aspect deeply understands a team's visual (video, 3D, multimodal docs) libraries and day-to-day workflows, performing clip search, rough drafting, and automated review. For research: Aspect processes dense multimodal datasets, providing tools for segmentation, QC, structured schema extraction. Companies today have access to Glean and Notion Agents for working with vast text corpuses - Aspect is enabling that for multimodal data.

PlayVision is the AI Moneyball for Sports: the future for scouting, recruiting, and playing the game. Basketball teams upload their practices or games, and PlayVision uses AI and computer vision to automatically tag and analyze the film, giving coaches and scouts 10 times the information on their players and their opponents. PlayVision also uses algorithms to rank every single college and professional basketball player in the country with over 40 different metrics from tracking and play-by-play data. Check out our website at withplayvision.ai, and follow our instagram @playvisionai!

We help oncologists and radiologists extract insights from CT scans to support tumor characterization and treatment. The hospitals we work with include Stanford Hospital, Cedars-Sinai (the largest hospital in California), UCI Health, and Weill Cornell.

Allus builds next-gen vision foundation models that bring real intelligence to manufacturing. Enabling factories to see, understand, and improve production in real time.

Structured AI is building the AI workforce for construction design engineering, starting with AI agents that perform quality control on technical documents and drawings. Our agents learn your standards, apply the right building codes, and automatically review your drawings, so you see fewer clashes, fewer inconsistencies, and fewer costly RFIs and change orders. ⚙️ Automates QA/QC across mechanical, electrical, structural and more 🧩 Used by multi-disciplinary design firms to deliver projects faster with less risk 🚀 Mission: Free engineers to focus on creative design, not repetitive admin

OnDeck is the infrastructure layer that makes Vision Language Models accessible and scalable for enterprise. We let organizations instantly find any object, behavior or event, in any footage, without needing to train a model or collect any training data. The Pain: Creating vision models usually takes months: collecting training data, training, then deployment. Worse yet: + it’s often impossible to get enough data for a specific task, and + even the best cv models struggle to generalize across diverse camera setups, workflows and environments. To overcome these blockers, we bet early on the power of VLMs and built a vision engine that can generalize across any task and doesn’t need any training data. We published a NeurIPS workshop paper showing our new methods with VLMs beat traditional CV even at niche tasks. Our current customers include: - National Defense Organizations - Robotics Research - Security cameras - Behaviour analysis for port monitoring - Off-shore oil & gas monitoring

Juxta is creating the first Universal Positioning System (UPS) — a GPS alternative that can monitor movement anywhere on earth with no new hardware, minimal internet connection, and remote deployment. Using custom machine learning models and a first-of-its-kind synthetic fingerprinting technology, Juxta can make any indoor, outdoor, or underground space location-aware in minutes for a fraction of the cost of any other geospatial technology. Juxta's UPS can be leveraged across industries — from securely tracking soldiers in combat zones to providing offline positioning for robots to offering asset tracking in warehouses and ports around the globe.

Kirana AI is building the first full-stack AI store manager. With 100 years of family experience owning and operating grocery stores, we are starting there to prove our impact. Our first product utilizes an on-premise GPU that monitors the store 24/7, scans for theft, looks for workplace health and safety issues, and optimizes customer service. With future POS/ordering data integrations, our manager will automate many of the day-to-day tasks of running a store, such as ordering, pricing, assortment, and delegating.

Code Four slashes the administrative burden on law enforcement through AI-driven automation. Officers spend up to a third of their shifts manually transcribing body camera footage, creating inaccuracies, reducing transparency, and cutting into frontline duties. We're here to eliminate these inefficiencies by integrating with body cameras, record management systems, and digital evidence management systems—acting as a software layer for data processing. We free the most overworked and underappreciated people in the world: police officers.

LineWise is building an AI assistant for manufacturing teams to operate and resolve production issues, from defects to breakdowns. LineWise automatically extracts frontline know-how from live video recordings, turning them into structured SOPs and training guides. It also ingests generated SOPs, machine manuals, and past breakdown logs to suggest the most probable root causes when issues occur. Operators receive step-by-step guided troubleshooting, and as they confirm outcomes, LineWise learns from each resolution - converting every fix into a reusable playbook and proactively recommending preventive actions to cut future downtime.

Weel is the next-generation all-in-one driver co-pilot that redefines your driving experience by turning your smartphone into a complete safety and navigation hub. Our free app offers a built-in dashcam with easy video sharing, a smart navigation system, and advanced AI features—such as augmented reality navigation and ADAS—that reduce your risk by up to 30%. Say goodbye to expensive dashcams and complex setups – with Weel, you get complete driving safety and guidance in one free smartphone app.

Exla aggressively quantizes AI models to minimize memory usage and maximize inference speed. Whether you're deploying LLMs, VLMs, VLAs, or custom models, Exla reduces memory footprint by up to 80% and accelerates inference by 3–20x - all with just a few lines of code. https://cal.com/exla-ai/schedule

Optifye.ai does AI performance monitoring for factory workers We place cameras in factories and use vision AI to tell supervisors who's working and who's not in real time. The shop floor has historically been a black box. With Optifye, manufacturing companies can now accurately measure worker output and boost efficiency!

City building departments are understaffed and overworked. Meanwhile, builders lobby to remove power from building departments. Because building departments are usually the only profitable/self-sustaining branch of a local government, this is problematic. Permitify allows city build-plan reviewers to 10x their productivity covering the labor shortage, reducing the time required to issue a building permit, and allowing the city government's only profitable branch to stay intact.

Mecha Health builds foundation models to automate x-ray analysis for radiologists. We take medical images and process them using proprietary models to produce accurate draft medical reports. Our first model was built in less than two months, and beat Microsoft, Google, and OpenAI on clinical accuracy metrics. On top of that, it’s two orders of magnitude smaller and trained with a quarter of the data. We are partnering with the largest privately owned radiology practice in the US and a multinational tele-radiology company to provide them with their own foundation model, enabling their radiologists to go from reading 1 scan per hour to 1 scan every 5 minutes. By charging on a per scan basis, x-ray report generation represents a 40B+ market opportunity.

Digitalcarbon transforms ordinary images and videos into interactive, photorealistic 3D environments. Our technology can take photos from any device and render 3D scenes at more than 100 frames per second on everyday devices. The founders were the first and second hires at a previous YC company AssemblyAI, and have more than 8 years of experience building AI models. Creating immersive 3D experiences typically requires complex equipment and suffers from slow rendering speeds. This has limited the adoption of 3D technology across various industries. Digitalcarbon fixes that, eliminating the need for specialized equipment and dramatically accelerating the rendering process. Our technology enables applications across various industries: * Real estate: Creating virtual property tours that feel remarkably lifelike * E-commerce: Providing interactive 3D product visualizations to boost consumer experience * Tourism: Helping tourist businesses attract more travelers with immersive virtual previews * Drone mapping and inspection: Enhancing aerial surveys and structural assessments with detailed 3D models Think Unreal Engine meets Matterport with no hardware, but faster, editable, and more accessible for businesses of all sizes.

With autarc we are building the OS for Europe's One-Stop Energy Installers. The problem: Energy installers are eager to expand their deployment of heat pump and photovoltaic systems, yet they are constrained by inefficient sales, planning, and installation processes. The solution: Our B2B software integrates CRM, planning, and design tools into a unified platform. This enables SMB installers to drastically cut down the pre-installation phase of home energy projects—from several weeks to just minutes. By leveraging LiDAR for spatial analysis, computer vision for automated assessment, and AI for system recommendations, we streamline the entire workflow. This results in a time reduction of up to 90%, enabling installers to confidently transition to and scale up sustainable home energy solutions.

Bucket Robotics is making defect detection faster, easier, and more deployable — starting with the 250M lbs of plastic wasted annually in U.S. manufacturing. Our platform turns CAD files into defect detectors. We generate synthetic, photorealistic training data to help factories catch flaws before they ship. No manual labeling, no real defects required. Our models deploy to edge hardware and integrate easily into existing automation stacks. We come from the self-driving world (Argo AI, Uber ATG, Stack AV), where we built reliable perception in high-noise, real-world environments. We're applying that experience to manufacturing: robust sensing, user-friendly interfaces, and fast iteration cycles. Manufacturers hesitate to adopt new sensing due to data concerns, integration risk, and poor UX. We’ve handled petabytes of regulated autonomy data — and built systems that earn trust. Legacy vendors like Keyence and FLIR offer hardware-centric tools with bloated pricing and outdated software. We’re building the opposite: flexible, modern tools that engineers want to use. Manufacturing is in the middle of a $700B automation wave across North America. Bucket Robotics is building the quality control infrastructure to match.

Open-source software for detecting AI-generated images (diffusion, GAN, deepfakes).

EdgeTrace unifies your entire camera network into a single, powerful intelligence platform–helping public safety teams detect threats faster, respond in real time, and protect the community.

DeepNight is building the next generation of night vision with AI.

Ocular AI is the data annotation engine for Generative AI, Computer Vision, and Enterprise AI models. We help you transform unstructured, multi-modal data into golden datasets to power generative AI, frontier models, and computer vision. Ocular Foundry is the most intuitive, data-centric, and fastest platform that lets you label, annotate, version, and deploy your data for training models. It also orchestrates your annotation jobs, improving collaboration with members and annotators. With Ocular Bolt, shift from humans in the loop to experts in the loop to supercharge your data labeling and annotation projects. Our global expert workforce ensures fast, accurate results—no matter the scale or complexity of your data. Companies spend huge amounts on training data, but Foundry and Bolt are AI-native tools that lower costs, reduce manual effort, and accelerate high-quality data collection. We’re replacing outdated, clunky, and expensive data software!

Dragoneye helps devs build their custom video detection models in less than 5 minutes. These models can then be used to power apps and features that use images and videos. No more arduous process of annotating any training data or doing any machine learning work themselves. Try out our demo today at https://playground.dragoneye.ai/ !

We are combining synthetic biology and artificial intelligence to make cancer detection as easy as a pregnancy test. Cleancard's mission is to bring lab-grade diagnostics into the comfort of your home. Our novel method enables robust diagnostics and biomarker tracking from the urine. We are currently developing fully at-home, 30-minute tests for cancer using this methodology. We are constantly expanding the number of conditions we can detect from a single sample with our platform technology.

Make data-driven decisions. Optimize property management. With the property Digital Twin from sizeless, you can access highly accurate data to understand your property's details, make informed decisions on inspections, and work on the required management optimization measures.

Shasta Health builds AI agents to handle patient calls, verify insurance and automate admin tasks for healthcare clinics

MICSI is introducing AI software that doubles resolution and halves scan time. This breakthrough enables imaging centers to significantly enhance their capacity and patient throughput, potentially saving countless lives and generating an additional $2 million of revenue per MRI scanner. Our initial offering serves as a stepping stone toward the company’s larger vision of transforming the MRI into a truly quantitative instrument that is capable of providing highly reproducible data for more accurate diagnoses and patient management.

Dream3D builds generative models to simulate and emulate worlds, real and imagined.

The DROP by Rex.Fit is an AI wearable camera that automates nutrition tracking using computer vision. It has garnered more than 93k USD in preorders (~450 units) since its launch last month. The nutrition estimation and recommendation tech behind the DROP is already integrated as an API for fitness chains, generating hundreds of thousands of dollars in yearly revenue for them. The DROP is built by brothers, sports nutritionists and computer vision engineers from the Swiss Federal Institute of Technology (ETH Zurich).

Scanbase makes it easy for medical companies to convert photos of rapid diagnostic tests into results. We do this by providing a simple API that any medical company can access.

Cerrion helps manufacturers automatically detect, understand and eliminate problems on their production lines using video-based Computer Vision. Our AI leverages standard CCTV cameras and learns how a manufacturing process looks like when things are going well and can automatically detect and track problems in real-time. For example, one of our customers, a Pepsi supplier producing 500 bottles per minute now automatically detects and reacts to a fallen bottle before it starts blocking their production line.

Apply Design is a leading AI-powered virtual staging app that empowers real estate professionals to showcase the full potential of their properties to prospective buyers, helping them sell homes faster and for higher prices. We virtually stage over 10,000 properties monthly, transforming images of vacant or outdated spaces into fully furnished, captivating property photos. Our proprietary AI-powered software is the only solution capable of generating photorealistic designs in one click while also enabling instant customization of any detail with complete control. This is a $10B market opportunity, given the 50M properties sold or rented annually in North America and Europe alone.

Every breakthrough AI application, from foundation models to autonomous vehicles, relies on processing massive volumes of images, video, and complex data. But today’s data platforms (like Databricks and Snowflake) are built on top of tools made for spreadsheet-like analytics, not the petabytes of multimodal data that power AI. As a result, teams waste months on brittle infrastructure instead of conducting research and building their core product. Eventual was founded in 2022 to solve this. Our mission is to make querying any kind of data, images, video, audio, text, as intuitive as working with tables, and powerful enough to scale to production workloads. Our open-source engine, Daft, is purpose-built for real-world AI systems: coordinating with external APIs, managing GPU clusters, and handling failures that traditional engines can’t. Daft already powers critical workloads at companies like Amazon, Mobileye, Together AI, and CloudKitchens. We’ve assembled a world-class team from Databricks, AWS, Nvidia, Pinecone, GitHub Copilot, Tesla, and more, quadrupling our size within a year. With backing from Y Combinator, Caffeinated Capital, Array.vc, and top angels from the co-founders of Databricks and Perplexity, we’re looking to double the team now. Join us—Eventual is just getting started. Please note we are looking for someone who is willing and able to come into our San Francisco office in the Mission district 4 days / week.

Tenyks (YC S21) is a Cambridge University spin-out building the World's Most Versatile Visual Intelligence Platform—think Snowflake for Vision or GPT for Video. Problem: Companies today collect oceans of visual data (from cameras, drones, and satellites), but struggle to extract actionable insights. Current solutions rely heavily on manual analysis by call centres in locations like the Philippines or AI platforms narrowly focused on security rather than deeper operational insights. This leaves critical improvements in efficiency, customer satisfaction, and revenue growth unrealised. Solution: Tenyks unlocks unprecedented insights by applying advanced multimodal Vision-Language Models (VLMs) capable of comprehensive video summarisation, retrieval, and visual question-answering. These VLMs are seamlessly combined with specialised analytics for emotion detection, activity recognition, object tracking, and even gait analysis. Our platform transforms ordinary visual footage into extraordinary operational intelligence. Initial Focus: Starting with Quick Service Restaurants (QSRs), we enable operations leaders to instantly understand customer flow, kitchen efficiency, staff productivity, and quality of service using their existing security camera footage—turning previously inaccessible data into direct operational improvements and clear financial ROI. For instance, insights from Tenyks can accelerate the speed of service significantly; even a 1-second improvement in service speed could generate $30,000 additional annual revenue per restaurant. For a multi-unit owner of 100 restaurants, that's an additional $3 million per year. Traction: We've recently secured our first Fortune 500 customer, joining other major enterprises, global system integrators, and high-growth startups trusting Tenyks for powerful, data-driven decision-making. Why Tenyks: Our patented technology originates from the founders' PhD research at Cambridge University. Previously, we helped world-leading teams scale high-performing Vision AI in production. Recognition: Recognised as a Rising Star in Gen AI by Sifted, Best Product of the Year by Vision AI Alliance, and Company of the Year by Cambridge University (previously awarded to DeepMind).

Mach9 is at the forefront of leveraging advanced machine learning and computer vision techniques to transform geospatial data into actionable insights for urban development and infrastructure management. Our cutting-edge technology automates the detection and analysis of key urban features, aiding in the planning and maintenance of transportation networks.

SnapCalorie is the world’s first app where you can take a photo of ANY meal and get an accurate calorie count. It takes seconds to use and is more accurate than a trained nutritionist. The team previously co-founded Google Lens, Cloud Vision API, and published a CVPR paper demonstrating the first algorithm to outperform a professional nutritionist at calorie counting.

MarqVision’s mission is to protect and build a future shaped by original ideas, innovations, and creativity. As a leading online brand protection solution, we help global brands identify and remove counterfeits and combat piracy from over 1,500 platforms and thousands of rogue websites in over 118 countries. In a time when counterfeiting poses an increasingly serious challenge, we believe that democratizing legal expertise and access to the latest in generative AI technology is critical for continuing to support the inventors, developers, and artists of the world who work tirelessly to deliver safe, high-quality products. MarqVision’s technology powers everything from detection, monitoring, and even enforcement to protect brands at scale. Founded in 2020 by Harvard Law graduates, MarqVision is proudly backed by Altos Ventures, DST Global Partners, Softbank, Atinum Investments and Y Combinator. Our commitment to innovation has also been recognized with a prestigious 2022 Innovation Award from LVMH Louis Vuitton Moët Hennessy (LVMH), and we are honored to be part of the LVMH accelerator program, La Maison des Startups, at the Station F incubator. As we bring forth the next evolution of brand protection, we invite businesses everywhere to join us in safeguarding the creativity that drives our world.

LightTwist is your virtual studio in the cloud. Record and stream video in a customizable virtual studio that you control from your browser.

Encord is the leading data platform for AI. We help top AI teams get models into production faster with human-in-the-loop workflows for labeling and RLHF, infrastructure for data-centric model testing, and powerful data curation and management.

Roboflow enables developers to make the world programmable. Use our tools to build better datasets (collect image, video / annotate), models (foundation and fine tuned small models), and deployments (self hosted, edge, APIs, SDKs) for computer vision. Over 250k developers, including those from over half the Fortune 100, build with our open source and hosted tools. Build with us: https://app.roboflow.com Hack with us: https://roboflow.slab.com/public/posts/roboflow-hackathons-external-u478m1iz Work with us: roboflow.com/careers

Hellometer helps fast food owners' grow same store sales using security cameras to improve customer wait times. Speed of service is fast food's core value proposition (it’s literally in the name) and owners currently don't have a reliable way to measure it. We show them when and where they have slowdowns and for every 7s they improve service times, owners see 1% increase in revenue.

Photoroom provides photo editing software powerful enough to create outstanding images yet simple enough to be used without any training. We leverage deep learning to translate pixels into objects, drastically simplifying non-creative tasks such as removing backgrounds from images. Our mission: enable entrepreneurs and small businesses to compose images that stand out.

Taiv puts TVs on autopilot, automatically switching between cable, our own streaming channels, digital signage, and trivia. Our proprietary AI model analyzes the live video feed and intelligently switches sources during commercial breaks, show changes, or based on the time of day, to show the best possible content & ads for any environment. We offer this for free to venues and make money by selling ads, which we rev-share back with the venue, turning their TVs from a cost center to a profit center.

Oda is an AI agent for home design. Homebuyers and renters alike can use Oda to discover their interior design preferences, apply those preferences to their new home, and then find the best furniture, décor, and appliances aggregated from many different online retailers. Oda distributes its product to homebuyers and renters through partnerships with real estate companies and platforms. Real estate partners add Oda AI iFrames to their websites to showcase listed homes with AI-generated designs. Each AI iFrame is directly linked to an Oda board, allowing consumers to transition from an iFrame to Oda to further personalize the designs and create their dream home.

We analyze thousands of video streams to find and track people without facial recognition. Our tech is available as an API and has multiple use cases. Unique people counting, forensic people search, falsa alarm filtering and many more.

Aura Vision is like Google Analytics for physical stores. We help retailers make more sales by using their existing security cameras and our plug & play AI to better measure their customers and stores.

Overview is a company that takes the cutting edge in computer vision and deep learning and applies it to previously unsolvable manufacturing inspection problems. We are truly a full stack company. We install physical cameras into the facility, run inference on the edge and manage massive deployments. Overview also streams gigabytes of video/image data to the cloud for our web platform to give customers advanced insights and analytics.