Skip to content

qFoldIT

LEGO brick road to open science on the cover of Trends in Biotechnology, September 2022. UEFN MCP

ANTHROPIC

qFoldIT — scientific computing, human-in-the-loop discovery and interactive runtimes

GAMEPLAY - LEGO FORTNITE SCIENCE/TSMC GAMEPLAY - LEGO FORTNITE SCIENCE GAMEPLAY - LEGO FORTNITE SCIENCE

qFoldIT — Science, Compiled into Missions

Contract-first scientific computing and human-in-the-loop discovery across MCP, UAG, simulation and interactive runtimes.

GitHub Organization · Public State & Project Site · IP & Evidence Registry · Canonical Architecture


What qFoldIT is

qFoldIT is a scientific mission platform that converts research and industrial objectives into governed, executable and auditable human/AI missions.

The platform separates scientific authority from presentation, gameplay, learning and commercial policy:

Scientific / Industrial Objective
        ↓
Mission Contract
        ↓
Capability Profile
        ↓
Mission Router
        ↓
Mission Compiler
        ↓
UAG / Scientific Object Model
        ↓
Runtime Adapter
        ↓
Human + AI Search
        ↓
Submission / Event
        ↓
CAMEO Reconciliation
        ↓
Scientific Validation
        ↓
Evidence / Contribution Record
        ↓
STATE Projection
        ↓
Enterprise Deliverable

The result is not merely a game level. It is a validated scientific workflow that can be rendered as an interactive experience when that is the best interface for the mission.

qFoldIT in Claude / MCP environments

qFoldIT is designed to operate as a scientific capability fabric for MCP-compatible agents, including Claude-based research workflows.

The correct boundary is:

Claude / Scientific Agent
        ↓
MCP Scientific Tools & Services
        ↓
qFoldIT Mission Contract
        ↓
UAG / Scientific Object Model
        ↓
Mission Compiler / Runtime Adapter
        ↓
Execution
        ↓
Scientific Validator

Claude is an orchestration and reasoning interface. Scientific truth remains external to the conversational layer and is established by the configured scientific validators and customer-authorized reference boundaries.

qFoldIT therefore does not claim that an LLM response is itself a scientific result, nor that an interactive runtime can redefine scientific authority.

Core architecture

Layer Function Representative qFoldIT surfaces
Governance Rights, provenance, contracts, evidence .github, .github-private, IP-valuation-registry
Enterprise Control Plane Customer objectives and mission lifecycle CORPORATE_APP
Compilation Routing and deterministic mission compilation CORPORATE_APP, conformance fixtures
Scientific Object Model Canonical scientific/world representation Scientific-Object-Schema, UAG
Scientific MCP Mesh Machine-readable tools and agents qFold-MCP, Protein-MCP, Protein-Design-MCP, Atomic-MCP, scientific skills
Validation Fabric Authoritative scientific scoring OPENSTRUCTURE, domain validators, quantum services
Reconciliation Runtime submission → scientific evidence CAMEO-REALTIME-VALIDATION, INDUSTRIAL-CAMEO
Runtime Mesh Engine/browser/desktop execution UEFN, Unity, UNIGINE, Web, Standalone
Learning Strategy/coaching/compute optimization learning-policy, gameplay optimization, QOS
Attention / Presentation Participation and distribution original/licensed narrative and game layers
Commercial Policy Eligibility, rights, rewards and customer policy enterprise policy layer
STATE Safe public/organizational projection qfoldit.github.io

Canonical contracts

The organization is moving toward explicit, versioned boundaries rather than direct application-to-engine coupling.

qfoldit.mission/1.0
qfoldit.mission-routing/1.0
qfoldit.mission-compiled/1.0
qfoldit.uag/0.1
qfoldit.event/1.0
qfoldit.submission/1.0
qfoldit.evidence/1.0
qfoldit.learning-policy/1.0

These contracts let scientific services, enterprise applications, game engines, simulation environments and future runtimes evolve independently.

Runtime adapters

One scientific representation can be delivered through multiple interaction surfaces:

                    Canonical UAG / Scientific State
                                │
          ┌─────────────────────┼─────────────────────┐
          │                     │                     │
        UEFN                  Unity                 UNIGINE
          │                     │                     │
   UEFN-TOOLBELT         UNITY-TOOLBELT       UNIGINE-TOOLBELT
          │                     │                     │
          └─────────────────────┼─────────────────────┘
                                │
                          WEB-TOOLBELT
                                │
                         STANDALONE APP

Runtime adapters are execution surfaces, not scientific authorities. The Web runtime intentionally consumes canonical state rather than embedding heavyweight scientific solvers; the Standalone runtime provides an offline-first desktop experience with local interaction and MCP bridging. Existing adapter work explicitly tracks third-party licenses and attribution.

Scientific validation

qFoldIT's current OpenStructure integration establishes a server-side scientific validation boundary with protected references and native structural scoring.

CORPORATE_APP
      ↓ authenticated server-side request
OPENSTRUCTURE qFoldIT service
      ↓
OpenStructure / `ost`
      ↓
LDDT / local-LDDT / QS-score
      ↓
ValidationResult
      ↓
validation.completed + audit evidence

A failed validator does not become a score by fallback logic. This fail-closed rule is fundamental to the qFoldIT evidence model.

CAMEO and three clocks

qFoldIT distinguishes:

  1. Authoring clock — target creation and mission compilation.
  2. Runtime clock — interaction and event capture.
  3. Reconciliation clock — external scientific validation and evidence publication.

For UEFN this distinction is essential: live Verse execution is not treated as a direct arbitrary-network scientific client. The intended model is runtime interaction → session publication → external reconciliation → scientific validation.

Human + AI discovery

qFoldIT treats players and agents as search participants rather than as sources of scientific truth.

Mission State
+ Player / Agent State
+ Historical Outcomes
+ Compute State
        ↓
Learning Policy
        ↓
Decision / Coaching
        ↓
Runtime Outcome
        ↓
Validation
        ↓
Evidence
        ↓
Policy Update

The target metric is validated scientific contribution density, not raw playtime.

Attention Capture

Game design, narrative, cultural presentation and interactive distribution are first-class platform capabilities, but they remain downstream of mission and scientific authority:

Original / Licensed Presentation
        ↓
Quest Object
        ↓
Scientific Mission
        ↓
Human Search
        ↓
Validation
        ↓
Contribution

Third-party brands, entertainment franchises and external ecosystems remain governed by their own licenses and contracts. qFoldIT does not claim ownership of upstream or partner IP merely by integrating with it.

Current scientific mission domains

qFoldIT's mission architecture is designed for:

  • protein and peptide design;
  • structural biology and molecular modeling;
  • drug discovery and ADMET;
  • synthetic biology and L-systems;
  • materials and crystal search;
  • batteries and energy storage;
  • semiconductors and AI hardware;
  • quantum computing and hybrid optimization;
  • aerospace and space science;
  • energy, hydrogen and CCUS;
  • mining and resource processing;
  • industrial engineering and process optimization;
  • scientific software and cloud AI.

Evidence discipline

qFoldIT distinguishes:

  • E1 — verified implementation evidence;
  • E2 — strong architecture/product evidence;
  • E3 — design, roadmap or option evidence;
  • EX — third-party/upstream evidence.

Repository size, external ecosystem reach, prospect lists, TAM, gameplay hypotheses and third-party scientific achievements are not automatically qFoldIT-owned value or revenue evidence.

Visual and interaction heritage

LEGO brick road to open science

Baker Lab / qFoldIT visualisation

qFoldIT UEFN Toolbelt

LEGO Fortnite Science

Explore

Contact

qFoldIT
qFoldIT@gmail.com
https://github.com/qfoldit

Built on the open Model Context Protocol (MCP) standard.

Scientific authority remains with validated scientific services and authorized reference boundaries.

Historical profile material

The previous organization profile, including its complete historical image set, contact information, links, MPNC material and presentation examples, is retained verbatim in profile/README-LEGACY-2026-08-21.MD.

Pinned Loading

  1. QuPepFold QuPepFold Public

    Python 1

Repositories

Showing 10 of 32 repositories

Top languages

Loading…

Most used topics

Loading…