01 / Architecture
Control the cost of growing capability.
ERAIS separates the repertoire available to a system from the work selected for a task. We’re measuring how that affects quality and resource use as the repertoire grows.
Research, products and partnerships / Queensland, Australia
Fund the next stage of reusable intelligence.
NeuroForge is an Australian AI lab developing ERAIS and Fracture. We want AI to keep useful skills, add new ones and put the right amount of work into a task. Funding would help us test that approach and build useful tools on practical hardware.
We’re looking for partners to help test the technology and bring useful products into use. We haven’t published a fixed raise amount. Public information only; not an offer to issue securities.
NEUROFORGEIO PTY LTDACN 701 314 329 · ABN 87 701 314 329
No fixed askcurrent raise amount not published
Evidence-ledresearch and delivery milestones
Pre-revenuefounder-led development
The opportunity
Turn the research into tools people can use.
We’re developing ERAIS and Fracture alongside practical tools and services. We agree the scope and checks for research and deployment projects before starting.
02 / Transition
Bring useful existing capability with us.
Fracture has converted speech components and complete supported Minecraft checkpoints into native ERAIS execution. Recorded replays agree with full-compute references, and learned routing updates survive a restart. Quality, efficiency and deployment suitability still need verification for each workload.
03 / The company
Build a focused technical team.
Aligned support can add specialist research and engineering capacity, independent review and product delivery against defined evidence and customer milestones.
Want to help the experiments continue? Choose one-time or monthly support for the lab. This is separate from investment and does not buy shares or game access.
Native training, supported conversions, live experiments including Minecraft AI, and a public record of results.
The Discord source includes a knowledge mode with source and permission checks before delivery. A public native API completion and an installed Discord adapter response have been recorded on the shared ERAIS service. Knowledge-mode deployment, acceptance in a newly installed guild and signed Discord delivery are still pending.
QSEG27286 and SubjectNest’s 2027 Tools Competition Phase I entry were submitted on 29 September 2026; outcomes are pending. 77 Partners declined the Spring 2026 application and invited an Autumn 2027 update. CQU research scope and terms remain under discussion. None is an award or approved study.
No paid pilot, formal research partnership, award, grant or independent technical validation is claimed today. Current reviews remain scoped to their agreed questions.
The founder
Meet Lloyd.
Lloyd Handyside draws on biology, physics, chemistry, big data, programming, industrial work and academic research exposure.
That range of experience shaped ERAIS. Lloyd uses AI across research, implementation and testing while setting the technical direction and evaluating the results.
Read the founder storyPotential milestone areas
What funding would help us finish.
Funding would support specialist work, infrastructure and delivery. The research checks, product readiness and customer milestones would be agreed before committing the budget.
Test ERAIS on real work.
Measure quality and the memory and processing each task uses. Check that conversions preserve the original model’s numerical results, and arrange independent research review.
Put useful tools in customers’ hands.
Turn approved products and workflows into repeatable paid projects. Have customers check the results and keep evidence of what worked.
Bring supported models into ERAIS.
Use Fracture for supported conversions only. Each model and the service that runs it must pass the agreed tests.
Build a team that can keep it running.
Bring in research, security and operations specialists for agreed milestones. Make sure another operator can take over.
How the next stage will be checked
Current sequence
What the next stage needs to resolve.
We need to check what conversions preserve, close quality gaps and finish the checks for supported products and partner setups. Those results will guide what we build or offer next.
NeuroForge is the lab; ERAIS is the shared architecture; Fracture converts supported existing models into capabilities ERAIS can use. Consumer products, repeatable smaller-business work, and larger research or deployment agreements remain distinct commercial routes.
Research
Resolve the numerical-preservation and quality questions with frozen protocols and matched evidence.
Native ERAIS
Measure quality, resource use and continuation before making broader architecture claims.
Paid delivery
Turn suitable workflows into bounded work with real buyer decisions and recorded delivery effort.
Supported setups
Complete security, lifecycle, installation, recovery and operator gates for each supported configuration.
Next decision
Choose a funding, revenue, redesign or stop path from current evidence and delivery capacity.
Who we’d like to hear from
- Early-stage investors who see the value of AI that fits real operating constraints
- Deep-technology investors comfortable with explicit technical gates and controlled disclosure of proprietary work
- Industry partners with one valuable, constrained AI application and a person empowered to decide
- Research or compute partners who accept negative results and independent review
How we work together
- Agree the technical question, budget and decision before the experiment
- Retain successful, failed and inconclusive outcomes
- Arrange appropriate access for review of proprietary work
- Define delivery, ownership and acceptance in writing
Capital and research partnerships
Talk through what you have in mind.
Tell us what you’d like to fund or investigate, and what you would need to see before deciding. We’ll come back with what we think we can do and how we would check the result.