Capicú
Capicú is an early-stage edge AI company that builds CEMI, an open-source CLI tool enabling developers to shrink, test, and deploy ML models on edge hardware for biomedical and industrial applications.
- Company typePrivate
- Founded2025
- HeadquartersMayaguez, Puerto Rico
- Headcount1–10
- GTM typeB2B
- OfferingSoftware
What Capicú does
Capicú is an early-stage edge artificial intelligence company founded in 2025 and headquartered in Mayagüez, Puerto Rico, operating as Capicú Technologies LLC. The company develops CEMI (Capicú Edge ML Inference), an open-source command-line tool that enables developers to shrink, test, and deploy machine learning models directly on edge hardware where data is collected. CEMI provides a repeatable workflow for comparing models, runtimes, and deployment targets, explicitly addressing inconsistent model behavior following over-the-air updates, device variations, and compression artifacts — a problem Capicú describes as the "hard middle" between powerful models and practical deployments.
The company's applied focus is biology, healthcare, and industrial robotics, evidenced by two published case studies: edge medical vision for anemia screening (computer vision) and robotic arm fault monitoring and prediction (anomaly detection on time-series/structured data). Capicú distributes CEMI as a free open-source tool via pip and GitHub (github.com/capicu-pr/cemi, at v0.1.211 with 2,111 stars), with enterprise demand generated through direct outreach CTAs. The company has not disclosed pricing tiers, enterprise contract terms, or paying customer logos; revenue mechanics appear to rely on enterprise support, professional services, and custom implementation layered on the open-source core.
Capicú is led by CEO Sebastian A. Cruz Romero (computer scientist and bioengineer with 20+ fellowships, awards, and publications in ML and biomedical AI) and Head of Product Shenied E. Maldonado Guerra (coordinated 8+ engineers/operators in biotech manufacturing). The company has secured non-equity program support from AWS Startups, NVIDIA Inception, NSF ACCESS, and VentureWell, and collaborates academically with the University of Florida, Johns Hopkins University, and Universidad de Puerto Rico Mayagüez. Aggregate disclosed funding is approximately $60,000 across four small grants between December 2025 and May 2026.
Capicú firmographics
Firmographics- Name
- Capicú
- Legal name
- Capicú Technologies LLC
- Website
- https://capicu.ai
- Company type
- Private
- Founded year
- 2025
- Operating status
- Operating
- Headcount range
- 1–10 employees
- Short description
- Capicú is an early-stage edge AI company that builds CEMI, an open-source CLI tool enabling developers to shrink, test, and deploy ML models on edge hardware for biomedical and industrial applications.
- Ownership category
- akta.pro rank
Capicú industry classification
Industry- Product category
- Edge AI Deployment Software
- NAICS
- Custom Computer Programming Services (541511), Other Computer Related Services (541519)
- SIC
- Services-Prepackaged Software (7372)
- akta.pro primary industry
- Edge AI Benchmarking, Profiling & Monitoring (device telemetry, drift) (HDAAAJAJ)
- akta.pro secondary industries
- Edge AI Model Optimization & Compression (quantization, pruning, distillation) (HDAAAJAA), Model Deployment, Serving & Inference Platforms (HDAAABAF), Edge AI Deployment, Orchestration & OTA Model Updates (HDAAAJAI), On-Device Inference Runtimes & SDKs (mobile/embedded) (HDAAAJAB)
Keywords
Where Capicú is headquartered
LocationHeadquarters
- HQ city
- Mayaguez
- HQ country
- Puerto Rico
- HQ region
- Latin America
Offices1 record
Markets served
Capicú business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Marketing or Sales, Operations, Infrastructure
Revenue model
- Enterprise Support & Professional Services: Open-source CEMI tool available freely; enterprise engagements likely driven by support contracts, custom implementation, and professional services for organizations deploying at scale
- Open Source Core (Community): Free open-source tool (CEMI) distributed via pip/GitHub enabling broad developer adoption and community building around edge ML deployment workflow
Go-to-market motion1 record
Distribution channels2 records
Marketing channels4 records
Capicú product offering
Product offeringCore offering
Capicú builds and distributes CEMI (Capicú Edge ML Inference), an open-source command-line tool that enables developers to shrink, test, and trust machine learning models on edge hardware. The product provides a single repeatable workflow to compare models, runtimes, and deployment targets across over-the-air updates, runtime differences, device variations, and compression artifacts. Target applications include biomedical edge AI, medical vision screening, and industrial robotics fault monitoring.
Product overview
Capicú offers a single unified open-source product called CEMI (Capicú Edge ML Inference) designed to solve edge AI deployment challenges. The platform enables developers to shrink, benchmark, and deploy machine learning models directly on edge hardware where data is collected. The ecosystem includes case studies in medical vision (anemia screening) and industrial monitoring (robotic arm fault prediction), demonstrating the technology's application across bioscience and healthcare domains.
Differentiator
Problem solved
Functional benefit
Brands
- CEMI (Capicú Edge ML Inference): Open-source tool that helps developers compare models, runtimes, and deployment targets in one repeatable workflow for edge AI deployment.
Products and services
- CEMI (Capicú Edge ML Inference) Open-source command-line tool that helps developers shrink, test, and trust ML models on edge hardware. Enables a repeatable workflow to compare models, runtimes, and deployment targets, addressing edge deployment failures from compounding runtime, device, and compression differences. For ML engineers and developers building AI systems that run on edge devices.
Quantifiable outcome
- Repeatable edge deployment benchmarks eliminate surprises from over-the-air updates and runtime differences
Companies that use Capicú
Customer profileNamed customers2 records
Segments2 records
Ideal customer profiles2 records
Capicú technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
AI capability5 records
Feature3 records
Capicú partnerships and signals
Strategic signalPartnerships
Five partnerships are on record, tiered core.
- AWS StartupscoreAWS Startups program providing cloud infrastructure, credits, and go-to-market support for early-stage AI companies building on AWS.
- NVIDIA InceptioncoreNVIDIA Inception program for AI startups providing access to NVIDIA technology, hardware, and co-marketing opportunities within the GPU/AI ecosystem.
- University of FloridacoreAcademic research partnership contributing to Capicú's applied R&D in machine learning, bioengineering, and data science for biomedical edge AI applications.
- Johns Hopkins UniversitycoreAcademic partnership providing research collaboration and validation for Capicú's edge AI solutions in biomedical and healthcare domains.
- Universidad de Puerto Rico MayagüzcoreAcademic partnership contributing research capabilities and talent pipeline for Capicú's edge AI development in biomedical applications.
Scale indicators2 records
Recent moves6 records
Expansion highlights5 records
Capicú competitors and assessment
Company assessmentEmerging players
- Modzy: Modzy provides an enterprise platform for deploying, monitoring, and managing AI models at the edge and in production, including benchmarking and runtime comparison — overlapping with Capicú's repeatable benchmarking workflow at the enterprise tier.
- Landing AI: Landing AI focuses on visual AI for industrial and healthcare applications, with tools for building, deploying, and monitoring edge vision models. It overlaps with Capicú's vertical-specific edge vision focus but targets enterprise customers more directly.
Broad incumbents
- NVIDIA (TAO / TensorRT): NVIDIA offers TAO, TensorRT, and Jetson as a broad edge AI stack with model optimization, benchmarking, and OTA capabilities. Capicú's CEMI competes with this integrated offering, but as a niche, hardware-agnostic alternative.
- Hugging Face Optimum: Hugging Face's Optimum and inference tooling offer model optimization and benchmarking across runtimes and devices. It is a much broader ML platform but addresses overlapping developer workflows for on-device deployment.
- Qualcomm AI Hub: Qualcomm's AI Hub provides model deployment and benchmarking across Snapdragon edge devices, overlapping with Capicú's runtime/device benchmarking workflow as part of a broader mobile/edge silicon portfolio.
- Google MediaPipe / LiteRT: Google's MediaPipe and LiteRT (formerly TFLite) provide on-device inference runtimes and deployment tooling with benchmarking across devices, positioning as a broad incumbent for cross-platform edge ML workflows that Capicú's CEMI also targets.
Direct peers
- Roboflow: Roboflow is a developer-focused computer vision platform with strong PLG adoption and biomedical/industrial use cases. Like Capicú, it combines open-source tooling with enterprise sales and targets edge vision workflows.
- Deeplite: Deeplite provides AI model optimization and automation for edge devices, focused on compression, quantization, and runtime selection — closely paralleling Capicú's 'shrink, test, and trust' workflow for edge ML deployment.
- Edge Impulse: Edge Impulse is a leading edge AI development platform for embedded ML, now part of Qualcomm. It directly overlaps with Capicú's CEMI on benchmarking, deployment, and OTA workflows for on-device AI, serving similar developer and industrial personas.
- OctoAI (formerly OctoML): OctoAI optimizes and deploys ML models across hardware targets and runtimes, addressing the same 'hard middle' between model and deployment that Capicú targets. Now vertically integrated into NVIDIA's ecosystem, it remains a direct workflow competitor.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat4 records
Key risks7 records
Key highlights6 records
Customer concentration
Capicú social profiles
Digital presenceCapicú financial estimates
Financial estimateRevenue estimate
Valuation estimate
Capicú leadership team
Management profileNumber of profiles
Profiles2 records
Capicú funding detail
Funding detailFunding overview
Funding rounds4 records
Investors2 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Capicú M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Capicú
What does Capicú do?
Capicú builds and distributes CEMI (Capicú Edge ML Inference), an open-source command-line tool that enables developers to shrink, test, and trust machine learning models on edge hardware. The product provides a single repeatable workflow to compare models, runtimes, and deployment targets across over-the-air updates, runtime differences, device variations, and compression artifacts. Target applications include biomedical edge AI, medical vision screening, and industrial robotics fault monitoring.
Is Capicú a public or private company?
Capicú is a private company. It is classified as founder individual operated bootstrapped and is currently operating.
When was Capicú founded?
Capicú was founded in 2025. It employs 1 to 10 people.
Where is Capicú based?
Capicú is headquartered in Mayaguez, Puerto Rico, in the Latin America region.
How does Capicú make money?
Two revenue lines are on record. Enterprise Support & Professional Services are the primary driver. The others are open Source Core (Community).
Who are Capicú's main competitors?
Emerging players on record are Modzy and Landing AI. Broad incumbents are NVIDIA (TAO / TensorRT), Hugging Face Optimum, Qualcomm AI Hub and Google MediaPipe / LiteRT. Direct peers are Roboflow, Deeplite, Edge Impulse and OctoAI (formerly OctoML).
Does Capicú have an API?
No public API is recorded for Capicú.
What industry is Capicú in?
Capicú's product category is Edge AI Deployment Software. Its primary akta.pro industry code is HDAAAJAJ, Edge AI Benchmarking, Profiling & Monitoring (device telemetry, drift), with a secondary code of HDAAAJAA, Edge AI Model Optimization & Compression (quantization, pruning, distillation). Its NAICS code is 541511 and its SIC code is 7372.