Browse categories
Explore
Fiverr Pro
English
$
USD
I will build a private ai document assistant that runs fully offline
Maneet Goyal
About this gig
Your documents never leave the machine. That is the whole point.
I build cross-platform desktop apps with Tauri and Rust that parse PDFs, scans, and office documents entirely offline, then let users search, summarize, and chat with them using a local LLM (Ollama, llama.cpp, or an embedded model). No cloud APIs, no per-token costs, no data leaving the device. Ideal for legal, healthcare, finance, and research teams with strict privacy needs.
I am a software engineer with 7+ years of experience and 5+ years shipping desktop apps used in the US, UK, and EU, including document-OCR and AI-assistant desktop products. Tauri apps ship as small, fast native binaries for Windows, macOS, and Linux.
What you get:
- Tauri 2 + Rust core with a clean React/TypeScript UI
- Offline parsing: PDF, DOCX, images (OCR), CSV
- Local LLM chat, summarization, and semantic search (embeddings + vector store on device)
- Installers for Windows and macOS, code signing and auto-update on the top tier
- Clean, documented code and a handover walkthrough
Every project starts with a short scoping call so we agree on exactly what ships. Message me before ordering and I will confirm fit within a day.
Get to know Maneet Goyal
Maneet Goyal
Desktop apps for Windows and Mac, built to ship
- FromIndia
- Member sinceJul 2025
- Avg. response time1 hour
- Last delivery1 month
Languages
English, Hindi
Hi, I'm Maneet. Seven years building software, 85+ projects delivered for clients in the US, UK and EU.
I build desktop apps for Windows and Mac from one codebase, using Electron, Tauri and Rust. New products, existing web apps turned into desktop apps, and privacy-first tools that run AI models fully offline.
I also handle the part most developers skip: code signing, Apple notarization, auto-updates and CI/CD. You get a real installer, not a folder of files.
Tell me what you're building and I'll tell you what I'd do first.
My Portfolio
FAQ
Will my documents or data ever be sent to the cloud?
No. Parsing, embeddings, and LLM inference all run on the user's machine. The only network calls, if any, are for downloading the model once and for app updates, and both can be disabled.
Which local LLMs do you work with?
Ollama and llama.cpp based models such as Llama 3, Mistral, Qwen, and Phi. I will recommend a model based on your users' typical hardware and quality needs.
What hardware do end users need?
A modern laptop with 8 to 16 GB RAM handles small quantized models well. For heavier models I design a graceful fallback (smaller model or clear system requirements screen).
Why Tauri instead of Electron?
Tauri apps are typically 10 to 30x smaller and use far less memory, since they use the OS webview and a Rust core. I have shipped both, and for offline AI tools Tauri's footprint leaves more RAM for the model.
Do I own the code?
Yes. Full source code and IP transfer on completion, with documentation.
Can you integrate with my existing backend later?
Yes. The app is structured so an optional sync or API layer can be added without reworking the offline core.
Can we talk before I order?
Please do. A short message about your documents and users lets me confirm scope and timeline within a day.
