Archive position — measured, not model output
0 likes on Devpost
2,264 of the 7,856 archived projects have more likes, and 5,592 share exactly 0 — so this project's #5,880 place in the like-ranked listing is a tie-break inside that group, not a ranking.
Projects (log scale)
Likes on Devpost. ▲ marks this project's group.
Show the figures
| Likes | Projects | Share of archive |
|---|---|---|
| 0 | 5,592 | 71.2% |
| 1 | 1,758 | 22.4% |
| 2 | 285 | 3.6% |
| 3–4 | 132 | 1.7% |
| 5–9 | 75 | 1.0% |
| 10+ | 14 | 0.2% |
Executive Summary
What the company appears to be
Pemi is a browser-based web application that uses local AI inference (YOLO model) to detect cats and dogs in real time from a user's camera feed. When a pet is detected, it records a short silent video clip and uploads it for analysis by an external algorithm. The result includes a mood interpretation and a selected frame from the recording. The system runs entirely in the browser with local detection, then uploads to S3 and sends to a Node.js backend for processing and algorithmic analysis.
What changed
This is a self-reported project submitted as part of the OpenAI 2026 hackathon. It was built using vanilla HTML/CSS/JS, ONNX Runtime WebAssembly, YOLO models, and browser APIs like MediaRecorder and Web Workers. The product flow includes local detection, recording, upload, transcoding, and algorithmic analysis.
Single most important open question
Is there any evidence of a commercial business model or traction beyond the hackathon submission? The description contains no data on revenue, customers, usage metrics, or monetization strategy.
What The Product Actually Is
The description states that Pemi is a browser-based application for detecting pets (cats and dogs) using local AI inference. It uses an ONNX model running in the browser to detect animals and records a five-second silent clip when a stable detection occurs. The video is uploaded directly to S3, converted to H.264 MP4 if needed, and then passed to an algorithm for analysis.
The system includes:
- Local YOLO-based object detection
- Browser-based recording using MediaRecorder
- SVG birds orbiting the detected pet
- Knowledge cards with facts about cats or dogs
- A Node.js backend for transcoding and algorithmic processing
It is described as a "local first-stage detection" system, meaning no video is uploaded until a stable detection occurs.
Evidence
- The author states: “Pemi uses local YOLO cat/dog inference (ONNX Runtime WebAssembly)”
- The author states: “The browser never receives the S3 presign-service key or algorithm-service credentials.”
- The author states: “The first uncached visit currently downloads an approximately 9.9 MB ONNX model and a 26.8 MB JSEP WASM binary.”
Inference This is a browser-based AI application that leverages local inference for pet detection, with backend processing for video transcoding and algorithmic analysis.
Positioning & Claim Evolution
The project's tagline is: _“Getting to know your furkids with Pemi!”_
It positions itself as an AI-powered tool to help users understand their pets better through visual interaction and interpretation.
The author describes the experience as:
- “Local first-stage detection”
- “Mobile camera controls”
- “Pet-following birds”
- “Knowledge cards”
These elements suggest a focus on emotional connection, accessibility, and user engagement rather than pure functionality or enterprise use.
Evidence
- The tagline: _“Getting to know your furkids with Pemi!”_
- The author describes the experience as “passive companionship without adding scores, tasks, or extra network requests.”
Inference The positioning appears to be centered on emotional engagement and pet care, not commercial or enterprise use.
Target Customer & ICP
The description does not explicitly define a target customer or ideal customer profile (ICP). However, the product is built for general users who own pets and want to interact with them through a browser-based experience.
It is designed for:
- Users with access to a camera
- Mobile or desktop browsers
- Pet owners interested in understanding their pet’s mood or behavior
Evidence
- The author states: “The interface, responsive layout, camera preview, local detector, recording, birds, and knowledge cards can be reviewed without production credentials.”
- The product is described as working on both mobile and desktop.
Inference It appears to target general pet owners who are tech-savvy enough to use a browser-based tool. No specific segment or persona is defined.
Business Model & Pricing Evidence
There is no evidence of a business model, pricing strategy, monetization plan, or revenue streams in the description.
Evidence
- The author states: “The browser never receives the S3 presign-service key or algorithm-service credentials.”
- No mention of subscription plans, usage fees, or paid features.
- No data on customer acquisition cost (CAC), lifetime value (LTV), or monetization strategy.
Inference No commercial business model is evident from the description. It appears to be a hackathon project with no stated monetization.
Technical & Delivery Signals
The system uses:
- ONNX Runtime WebAssembly for local inference
- YOLO models for cat/dog detection
- MediaRecorder API for video capture
- Web Workers and OffscreenCanvas for performance
- Node.js backend for transcoding and algorithmic analysis
- Amazon S3 for storage
- FFmpeg for format conversion
The architecture is described as:
- Browser-based local detection
- Upload to S3 with presigned URLs
- Backend processing via Node server
- Polling of algorithm results
Evidence
- The author states: “Pemi uses local YOLO cat/dog inference (ONNX Runtime WebAssembly)”
- The author states: “The browser never receives the S3 presign-service key or algorithm-service credentials.”
- The author states: “The Node server creates an asynchronous task, polls its result every three seconds…”
Inference It is a hybrid system with local processing and backend orchestration. It uses modern web technologies and is designed for performance on mobile.
Traction & Maturity Signals
There is no evidence of traction or maturity beyond the hackathon submission. No data on:
- Users
- Revenue
- Customer adoption
- Product usage metrics
- Iteration history or product evolution
Evidence
- The project was submitted to a hackathon.
- The author states: “This section is therefore a transparent reconstruction based on repository history, implementation details, code comments, the product owner's account that the recent iterations were Codex-assisted.”
Inference The project appears to be in early development or prototype stage. No evidence of traction or commercial adoption.
Competitive Context
There is no mention of competitors or market context in the description.
Evidence
- No reference to existing products or services in this space.
- No discussion of how Pemi compares to other pet AI tools or platforms.
Inference No competitive positioning or market analysis is evident. The project does not appear to be part of an established market or ecosystem.
Key Risks & Red Flags
Key risks and red flags include:
- No commercial model: No evidence of monetization, revenue, or customer base.
- Prototype nature: Submitted as a hackathon project; no indication of product maturity or traction.
- Technical complexity: Relies on browser-based inference and video processing — may not scale well or perform consistently.
- Limited scope: Focuses only on cats and dogs, with no indication of expansion plans.
- No data or metrics: No evidence of user behavior, retention, or performance.
Evidence
- The project is described as a hackathon submission.
- No mention of users, revenue, or adoption.
- The author states: “The browser never receives the S3 presign-service key or algorithm-service credentials.”
Inference This is a proof-of-concept with no commercial viability or traction.
Diligence Questions To Ask The Founders
- What is the intended business model for Pemi?
- Are there any plans to monetize this product beyond the hackathon?
- How does the algorithmic analysis work, and who owns it?
- Is there any data on user behavior or engagement?
- What are the technical limitations of the browser-based inference approach?
- Has the product been tested with real users beyond the development team?
- Are there plans to expand beyond cats and dogs?
Investment/Partnership Verdict
Not evidenced.
There is no evidence of a commercial business, traction, or investment-ready model. The project is described as a hackathon submission with no indication of monetization, user base, or product maturity.
Confidence Low This is a self-reported, unverified description of a prototype submitted to a hackathon. No data on revenue, customers, or commercial viability is provided.
Inference It is not a viable investment or partnership opportunity based on the information supplied.
Source
Submitted to the OpenAI 2026 hackathon on Devpost. Project home on DevPost.
The analysis above was generated by a language model from the project's own one-line description. It is not independent research and contains no verified traction, revenue or customer data.