AuraForge Foundry
AuraForge Foundry helps a business build an AI organization that mirrors its real organization: its units, roles, processes and knowledge, run by governed AI agents connected to the systems it already uses.
Underneath, it coordinates many AI agents from different providers so they can plan, deliberate and build software, under governance that the organisation controls.
Products
- In development
AuraForge Foundry Web
The corporate and product site.
Domain: auraforgefoundry.com
- In development
AuraForge Foundry App
The product experience.
Domain: auraforgefoundry.ai
Current status
AuraForge Foundry is in development. It has not been publicly released. The components and use cases below describe what we are building; details will change as development continues.
The platform
AuraForge Intelligence Engine (AFIE)
Turns a rough idea into a plan. The path is idea → brainstorm → multi-frontier deliberation → plan.
Several leading (“frontier”) AI models from different providers each propose an approach and then review each other’s proposals. The plan records the chosen approach, the alternatives considered and the open questions.
AuraForge Supervisory Engine (AFSE)
The governance layer. Every request to an outside AI provider passes through it first.
It checks the request against the organisation’s policies, for example which providers may be used, what kinds of data may be shared and how much may be spent. It also writes an audit record before any data leaves.
AuraForge Execution Engine (AFEE)
Carries out approved plans. Work is split into “lanes”: one AI agent working on one task.
Each lane runs in its own isolated container, so agents cannot interfere with each other or with the host system, and each lane’s output can be reviewed on its own.
Memory Fabric
A shared store of plans, decisions, reviews and outcomes from earlier runs.
The other components read from it, so a new run can reuse what earlier runs learned instead of starting from nothing.
How the parts fit together
Idea
YouDescribe a goal, a feature or a question in plain words.
Plan
AFIEBrainstorm options, have several AI models from different providers debate them, then write a plan.
Check
AFSECheck each outgoing request against your policies and record it before any data leaves.
Build
AFEERun AI agent "lanes" in isolated containers to carry out the approved plan.
Review
PeoplePeople review plans and results before they are used.
Remember
Memory FabricKeep decisions and results so the next run starts from what was learned.
What it is for
From idea to reviewed plan and code
Start with a short description of a product or feature. AuraForge Foundry produces a plan that people can review and change, then builds against the approved plan in isolated lanes.
Multi-AI research and second opinions
Ask the same question of several AI models from different providers and see where they agree, where they disagree and why, instead of relying on one answer.
Governed AI for small and medium companies
Smaller organisations often cannot build their own AI controls. AuraForge Foundry is designed to give them policy checks, provider choice, spending limits and an audit trail by default.