Hey guys -So we’re all excited to start playing with our tiiny devices. As more and more of us get our devices online, I wanted to share a few tips I’ve learned so far. My background: I work in a tech-adjacent field (engineering) but my only experience with coding is a few classes back in high school, so I am writing t

I used Tiiny to try to create a game dev pipeline.

Here’s what I tried:
I grabbed the Hermes agent from the store, and ran it with Qwen 3.8 27B loaded up. I then told Hermes that I have the MCP loaded in Godot, and to please connect it. Which it did, ran a test, and was able to access the tools. I then asked it to create

I've been using Tiiny AI to help me with my code. I work mainly in the Intellij IDE and I figured I would share how to setup Intellij to use Tiiny AI.

Steps to hook up intellij.

1) First grab your API key from the settings in Tiiny AI

2) Go to Intellij settings and search for AI assistant (Providers & API keys)

3) Select O

I wanted to keep all my flight records on one map, but the travel app I was using limits free users to 40 trips. So I built my own with Tiiny.

I used Kilo Code in Code mode and started with this prompt:

Build me a web page that keeps track of all my flights on a world map. Every time I add a trip, it draws a nice curved 

I used TiinyOS to generate two sets of chibi stickers: one based on a Shiba Inu and the other on a Siamese cat.

Both animals were easy to recognize, and the chibi look came through clearly.

Prompt 1

Based on a Shiba Inu, generate a set of chibi-style stickers with different facial expressions and gestures. Do not repeat t

If you have poked at the Vault and were not sure what it actually does, this is the plain version.

It is two things under one name. The first is a document store you can search by meaning: you drop files in, the device chunks them, an embedding model on the NPU turns them into vectors, and they land in a local database.

I keep most of my notes and learning materials in Obsidian. I wanted to use Tiiny with them without moving everything to another platform. I installed the Copilot plugin for Obsidian and added Tiiny through its model settings. Tiiny has an OpenAI-compatible API, so I entered the endpoint and can now use its local model

I started my morning by asking Tiiny for a quick rundown of the latest headlines.

Tiiny grouped the news into sections, summarized the main developments, and included its sources. I could scan the topics first, then decide which stories I wanted to read in more detail.

My favorite feature was the link preview. Hovering o

The short version is that almost none of the model runs on the CPU.

This was read off a unit rather than out of a doc. A chat model is served by a llama.cpp server, but the only weights sitting on the CPU side are the token embedding table. Every transformer layer and the output head run from a PowerInfer bundle compile

By the end of the week, my notes were a mix of short updates, half-finished thoughts, and reminders. They named colleagues, staging access, legal wording that wasn't public yet, and finance numbers. I gave them to Tiiny and asked for a report on completed work, work in progress, problems, next week's priorities, and he

The setup took more testing than I expected, but it's working now. Hermes can send a request to Tiiny and use the response.

This is still a small experiment. My next step is to try the same connection on local development tasks, document analysis, and a few personal workflows. That should show me where the setup is usef

A spec sheet does not tell you what a box does when you give it real work, so this is a set of numbers measured on one instead.

What is on the page: five models, 23 runs, measured between 15 August and 15 September. Every result carries the build, the host and the NPU cost it was measured with, because a figure taken of

The store lists 50 models now, and the thing nobody says up front is that the number beside each one is a budget rather than a spec.

Your box has 100 NPU units. Everything you keep resident spends some: a 35B chat model is around 50, an image model around 32, a text-to-speech voice around 7, a small embedding model 1. T

I wanted to know whether Tiiny could compare two documents that describe the same product from different angles, so I built a small test out of a PRD and a launch review for a fictional product, with deliberate disagreements planted in them. Both files stayed on my machine. I put them in a folder, pointed Tiiny at it,

The representative quotes were the most useful part of this test. They gave me a quick way to check whether Tiiny’s conclusions matched what the customer had said.

The record was not a clean transcript. It was 58 minutes and roughly 5,800 words, mixing the customer’s answers with my questions, my notes, an observer’s no

I use Tiiny to learn the things I want to understand, and I'm starting with agents.

Before I go deeper, I want to see how agents behave: what they're good at, where they struggle, and how I can work with them. With agents, Tiiny becomes a tool I can give tasks to, observe, and learn from.

I'm still early in the process,

I asked Tiiny to help me clean up my Mac.

At first, I didn't realize Tiiny had a Task mode. I was using Chat mode, so Tiiny could only explain what to check and give me a cleanup guide for large files, old downloads, app caches, and unused applications.

Then I discovered Task mode. With it enabled, Tiiny can use tools to