Real-world
Personal assistant
Concrete workflows, including the tradeoffs and limitations reported by their authors.
this collection
A year of using local models as a general life assistant
About a year into daily local use, this setup covers private coding projects and a broader life-assistant role: web search, journal access, and photo access.
Two DGX Sparks as a daily driver, ROI aside
This person knows Qwen 3.8 Flash Next on two DGX Sparks can't match frontier intelligence, and says that's not really the point for them.
Tried local for daily use, went back to cheap cloud subscriptions
A local model ran fine, but between electricity, a maxed-out machine, and cloud models costing a few euros a month, the math didn't work out.
A receipt-scanning pipeline that wakes the PC on its own
A phone app scans a receipt, a Synology NAS notices the new file and wakes the desktop, and a local model turns it into structured spending data — no manual step in between.
Email triage and a dictation-to-proper-writing tool, alongside a paid subscription
Local models here don't replace a cloud subscription so much as complement it — handling the smaller, more private slice of the workload.
Local is for tinkering; Claude still does the real work
On a VRAM-constrained 5070 Ti, this person keeps local strictly in the 'learning and playing around' lane, and routes actual work through Claude with local models as a validation step at most.
Running a local model daily, straight from a phone
One of the only mobile setups in the thread — used daily on a phone as an offline-capable stand-in for search and reference lookups.
A local assistant for todo lists, finances, and life planning
A personal assistant handles organization, financial tasks, document understanding, and web searches locally, while heavier work remains in the cloud.
Eight daily agent workflows, driven from Telegram on a Strix Halo
A full personal-operations stack runs on one 128GB Strix Halo machine, with several named agents reachable through a single Telegram chat and each handling a different part of daily life.
A private morning brief over Obsidian notes and a calendar
A local assistant reads the owner's Obsidian vault and calendar to produce a morning brief, a weekly review and reminders of abandoned goals and ideas.
A Jetson that reads scanned mail and keeps an inventory and budget
A rule-forwarded inbox feeds a local workflow on an Nvidia Jetson that ingests invoices, receipts and cards, tracks spending and builds a contact and inventory database.
An avatar assistant with persistent memory shared between desktop and Second Life
A retired teacher with no coding background built an animated avatar assistant on a local 9B model, then bridged its persistent memory into a virtual world so it inhabits two bodies.
A family finance tracker built with Hermes and an MCP
A weekend vibe-coded web app plus a Hermes agent tracks family finances, with an MCP for adding expenses and inputs arriving from forwarded emails, Telegram or Discord.
Persona profiles and knowledge modules matter more than tok/s
A local setup built around two persona profiles, GPT-OSS 120B and a Swift coding model, prioritises configuration and knowledge structure over raw speed.
Running a real agent on Qwen 3.5 9B on an M1 Pro MacBook (16GB)
A model-agnostic personal automation system switched from a cloud model to Qwen 3.5 9B on an M1 Pro MacBook with 16GB of unified memory, handled a real subset of agent work locally, and ran 0.8B/2B models on an iPhone.
Creative writing and a Discord bot on an 8GB RTX 4060
A hybrid personal-assistant setup for creative writing, quick searches and Discord bot experiments on an 8GB RTX 4060, where 16GB of system RAM is the real ceiling
A Hermes agent coordinating hobbies, inventories and price tracking
On an RTX 4070 and 128GB of RAM, a Hermes agent is being set up to query a private SQL database of hobby equipment and watch prices on the owner's behalf.
My personal JARVIS
Personal life/project management agent with persistent memory, reminders, smart home duties (timers lights etc), computer use.
A quiet local personal assistant on a two-machine setup
A contributor uses a quiet Qwen 3.5 9B assistant on an RTX 5060 laptop for planning, documentation, research and private questions, while a Radeon desktop handles heavier work.
An agent that plans a wedding from your inbox
The same low-key homelab setup also runs a semi-agentic wedding planner that reads email, keeps a running database and document trail, and answers questions about where things stand — plus a separate loop that just deletes marketing mail.
Local to get past cloud filters and nannying
Part of a hybrid RAG setup spanning a Mac Studio, a MacBook, GCP, and OpenRouter, with an unrestricted local model used specifically for questions frontier chat products decline to answer.