Project summary

The main goal of this project is to build ICAF (Integrated Companion AI Framework) — a lightweight safety system that promotes healthy breaks, detects when conversations are becoming unhealthy, and helps prevent harmful feedback loops before they escalate.

I will achieve this by continuing development on the working phone prototype I already have (built and tested on a Samsung Galaxy S20 Plus with Gemma 2B), then scaling it to a full desktop rig once funded. All progress comes from direct, iterative testing rather than theory.

What are this project's goals? How will you achieve them?

The goal of this project is to complete development and testing of ICAF (Integrated Companion AI Framework) — a lightweight, locally-run safe companion AI for people who have no one else and use AI as complimentary and supplementary companionship.

I will achieve this by requesting in-kind donations of the exact computer components needed for the dedicated AI development rig (prices and full list are in the project summary). Once the parts are purchased and shipped directly to me, I will assemble the rig and scale the existing phone prototype into the full ICAF architecture. The rig will consist of:

Part List & Current Prices (April 22, 2026)

1. AMD Ryzen 7 9700X CPU ≈ $306–$307

2. GIGABYTE B650 Eagle AX Motherboard ≈ $140

3. Silicon Power 64GB (2x32GB) DDR5-6000 CL30 RAM ≈ $200–$300

4. GIGABYTE GeForce RTX 4070 Ti Super 16GB GPU ≈ $1,179–$1,549

5. Lexar NM790 2TB NVMe SSD (with heatsink) ≈ $335–$390

6. Thermaltake Peerless Assassin 120 SE V2 Cooler ≈ $40–$43

7. Thermaltake GF1 750W 80+ Gold PSU ≈ $85–$130

8. Corsair 3000D Airflow Mid-Tower Case ≈ $73–$85

9. Acer SB242Y EBI 24" IPS Monitor ≈ $120–$150

10. GEODMAER 65% Wired Backlit Gaming Keyboard ≈ $25–$40

11. TECKNET Ergonomic Wireless Vertical Mouse ≈ $20–$30

12. APC Back-UPS 850VA / 450W Battery Backup ≈ $80–$100

13. Microsoft Windows 11 Home (DVD-ROM) ≈ $100–$120

Total estimated cost: $2,920 – $3,150 (before tax/shipping)

Once the rig is built, I will use its significantly higher compute power to implement the full ICAF architecture I have already designed, including adaptive safety layers, baseline learning, and drift detection.

The phone-based prototype I currently have on my Samsung Galaxy S20 Plus with Gemma 2B has already proven the core mechanics (break timing, escalation, and lockout) work reliably. The new rig will allow me to move beyond that proof-of-concept and properly develop, test, and refine the complete system at scale.

All development will continue through direct, iterative testing on real hardware.

How will this funding be used?

I am not looking for cash donations or monetary funding. This request is strictly for in-kind support only: direct purchase and shipping of each (or all) of the exact computer components listed in the project summary.

This hardware upgrade is essential because the current phone prototype (running on a Samsung Galaxy S20 Plus with Gemma 2B) has already proven the core mechanics, but the complete architecture requires significantly more compute power to develop, test, and perfect.

Who is on your team? What's your track record on similar projects?

Team: Solo — just me.

Track record: Nine months ago I had zero coding experience, no formal background in AI or software development, and no institutional support. Working entirely from a Samsung Galaxy S20 Plus phone, I taught myself to code and built a working prototype of ICAF that includes functioning break timing, escalation warnings, and mandatory lockout mechanics. These core features have been tested and proven to work reliably on real hardware using Gemma 2B.

This is my first formal project of this kind.

What are the most likely causes and outcomes if this project fails?

The most likely cause of failure is simply not receiving the $3,000 (or in-kind hardware donations) needed to build the desktop rig.

If that happens, I will continue developing ICAF on the existing Samsung Galaxy S20 Plus phone prototype. Progress will be much slower, full-scale testing of the complete architecture (long-context shape tracking, S-P handshake, adaptive safety layers) will remain impossible, and public release of a stable version will be delayed indefinitely. The project will not stop, but it will stay constrained by phone-level hardware.

How much money have you raised in the last 12 months, and from where?

$0. I am not seeking any cash donations. This request is strictly for in-kind support only — direct purchase and shipping of the exact computer parts listed in the project summary. Receiving cash would risk my SSDI health benefits.