Yash Jangid
Yash Jangid is an individual software developer, engineer, inventor, and founder—not a company—who builds useful, shareable software. His work spans developer tools, backend and realtime systems, GPU/AI pipelines, game development, and open-source projects; his site says his projects have reached 300k+ people.
At a Glance
- Individual computer users who need fast, private local file search
- Developers and technical users searching source code, logs, documents, and data
- AI-agent users who want local MCP access to indexed files
- People needing cross-platform file transfer without accounts or cloud storage
- +1 more
AI Tools by Yash Jangid
(1)
Kensaku
Open Source Local File Search
Discussions
No discussions yet
Be the first to start a discussion about Yash Jangid
Latest News
Kensaku v1.0.15 released, fixing Indra OCR crawl stalls and lock contention and improving extraction timeout and progress behavior.
Kensaku v1.0.14 released with an in-app updater, configurable update checks, issue reporting, OCR improvements, and ONNX Runtime fixes.
Yash published the Cobbic shutdown post-mortem, explaining why he turned off the AWS-hosted realtime social app after building and opening it to users.
Published an AI image-compression experiment reporting up to 20× smaller images using a neural-network autoencoder.
Products & Services
A free, MIT-licensed, local-first desktop file launcher and search engine. It searches filenames, document and code contents, images via OCR, and meaning locally; it also exposes the local index to compatible AI agents through MCP.
Video tools with an AI editing engine. The portfolio describes perception running at 500+ FPS on under 6GB of VRAM and lists GPU, TensorRT, YOLO, and FFmpeg as technologies.
A web-based, peer-to-peer file-transfer tool. Two devices connect with a flight code or QR code and transfer files directly over a WebRTC data channel without sign-up, cloud upload, file-size limits, or a per-byte charge.
A discontinued realtime social app with pages acting as places, a clean editor, realtime chat, search, and a feed designed to rank conversations rather than followers. It used AWS, Terraform, pub/sub, Typesense, and a MiniLM-based feed.
Market Position
Yash's projects compete by emphasizing focused, privacy-preserving, low-friction alternatives. Kensaku positions itself against filename launchers/search tools such as Everything, Listary, and PowerToys Run by adding local full-text, OCR, semantic, multilingual, and MCP search while remaining open-source and MIT-licensed; it also contrasts with Spotlight, Raycast, and Alfred on platform scope and pricing. AirDelivery positions itself as a cross-platform, no-account alternative to AirDrop, WeTransfer, Drive, WhatsApp, LocalSend, and PairDrop, with direct browser-to-browser transfer and no cloud storage.
Leadership
Founders
Yash Jangid
Computer science student and independent software developer/engineer. He describes work in tools, backend systems, DevOps, machine learning and AI, with earlier experimentation in game development and 3D graphics; he was a Software Engineer Intern at PhotoGPT from January to September 2026 and currently builds Captos and Kensaku.
Executive Team
Yash Jangid
Founder, software developer, engineer, and inventor
Independent builder and computer science student. His official portfolio identifies him as the person building Kensaku and Captos, and records his 2026 PhotoGPT software-engineering internship focused on backend systems and production infrastructure.
Founding Story
Yash presents his work as a series of self-directed projects intended to solve practical problems and be worth sharing. AirDelivery began because moving a video between a phone and laptop felt unnecessarily difficult due to compression, logins, timers, or ecosystem restrictions; Kensaku was built around the idea that people should be able to search not only filenames but the contents and meaning of local files without sending data to the cloud.
Business Model
Revenue Model
The publicly described projects are primarily free, open-source, or free-to-use tools. Kensaku is MIT-licensed and local-first; AirDelivery is presented as having no signup, upload, timer, size limit, or per-byte cost. No paid subscription, API pricing, or commercial revenue model is stated on the reviewed sources.
Target Markets
- Individual computer users who need fast, private local file search
- Developers and technical users searching source code, logs, documents, and data
- AI-agent users who want local MCP access to indexed files
- People needing cross-platform file transfer without accounts or cloud storage
- Teams and creators interested in AI video and GPU/vision tooling
- Private local search across files, documents, source code, screenshots, and OCR text
- Searching a personal file index by meaning or natural-language descriptions
- Giving local AI agents read-only access to a user's file index
- Fast cross-platform transfer of files between phones, laptops, and other modern browsers
- AI-assisted video editing and GPU-accelerated computer-vision pipelines
- Offline medical-imaging experimentation and technical graphics/physics simulations
- AirDelivery reports 20K+ users in its first month and 300K+ visitors to date; the author's write-up also reports 27K+ share sessions and terabytes transferred.
- The portfolio says Yash's projects have reached 300k+ people.