Prototype surfaces
Donor app, agent, dispatch API, researcher portal
Volunteer compute infrastructure for cancer research
Put unused powerbehind the cure.
Apoptosis is building nonprofit infrastructure that can connect idle CPU and GPU capacity with verified cancer research teams— safely, transparently, and at meaningful scale.
Technical prototype builtFounding partners soughtSecurity review next
Dispatch model / prototypePilot phase
NODE / OPT-INVolunteer GPUContributor controlled
APSecure dispatch
QUEUE / REVIEWEDResearch jobInstitution verified
Workload families
Molecular, genomic, docking, and model validation
Validation design
Independent execution before results are accepted
Day pilot target
A controlled institutional proof before public scale
Architecture and pilot targets—not live research impact claims.
Why apoptosis?
Apoptosis is the process that tells damaged cells to self-destruct. Cancer learns to ignore that signal.
This project exists to put more computing power behind the researchers working to understand and restore it.Built in memory of Kait Shannon.
The opportunity
Compute access should not decide which questions get asked.
Modern cancer research depends on large, parallel workloads. Institutional infrastructure is essential, but suitable exploratory work can still compete for finite capacity. Meanwhile, powerful personal hardware spends long stretches idle. Apoptosis is engineering a trusted bridge between the two.
More questions than capacity
Compute-intensive experiments can queue behind other institutional work, slowing iteration for smaller and early-stage teams.
Capable machines waiting to help
Modern consumer CPUs and GPUs can contribute meaningful parallel work when the owner is not using them.
Infrastructure with one purpose
A reviewed dispatch layer can turn voluntary capacity into complementary research infrastructure.
Prototype code today
The system already has a spine.
The current prototype models the full path from contributor control to research job submission. The next phase is not adding a prettier dashboard—it is validating the architecture under independent review and a real institutional workload.
Donor experienceReact donor app
Control stays with the contributor.
Node registration, live status, resource throttling, contribution history, and a visible pause control.
Local agentNode agent + Docker
Jobs run inside defined limits.
Heartbeat reporting, task receipt, isolated container execution, and fallback polling for resilient operation.
Dispatch coreAPI + PostgreSQL + Redis
Work is split, assigned, and checked.
Authenticated APIs, job queues, real-time events, task assignment, timeout handling, and redundant result comparison.
Researcher portalReact research portal
Access begins with verification.
Institutional applications, reviewed access, job submission, progress monitoring, and result retrieval.
One network, two communities
Contribution without complexity.
The contributor keeps control. The research institution keeps oversight. Apoptosis handles the secure coordination between them.
A contributor opts in
The donor chooses when the agent can run, how much capacity it may use, and when it should pause.
A verified job is prepared
An approved institution submits a divisible, container-compatible workload for technical and security review.
The network does the work
Jobs are split, assigned, monitored, and checked before validated results are returned to the research team.
For future compute donors
Your hardware. Your limits. Visible impact.
- Choose resource limits and pause at any time
- See the approved project your machine supports
- Public release follows independent security review
For research teams
Bring a workload that can travel.
- Divisible, container-compatible compute jobs
- Institutional verification and pilot scoping
- Reference outputs for result comparison
Trust before scale
Designed so confidence can be earned.
Volunteer hardware should never require blind trust, and researchers should never accept unverifiable output. The platform is being built around controls that can be inspected, tested, and improved before public launch.
Readiness gateIndependent security review, adversarial testing, reference-run validation, and pilot remediation are required before the public donor network opens.
Verified research access
Institutional identity checks and manual review are part of the access model before a research workload reaches the network.
Isolated execution
The architecture uses containerized jobs, resource limits, non-root execution, and network-isolated workloads on volunteer hardware.
Results checked twice
The pilot design assigns task chunks to independent nodes and compares outputs before results are accepted.
Inspectable donor software
The donor-side agent is intended for open release so contributors and reviewers can inspect what runs on a machine.
Security teams can request the prototype architecture and help define the independent review scope.
Review the security modelPilot workload families
Built for work that divides cleanly.
Final pilot eligibility depends on data sensitivity, container design, expected runtime, hardware variance, and independent technical review.
Molecular simulation
Protein-folding and molecular-dynamics workflows such as OpenMM and GROMACS.
↗Genomic alignment
Parallel sequence-comparison workflows using tools such as BLAST and BWA.
↗Drug docking
High-volume ligand-docking batches for early-stage compound screening.
↗Model validation
Divisible machine-learning training and validation jobs on approved datasets.
↗The first proof
A controlled pilot with measurable exit criteria.
The target is a 90-day institutional workload that can be compared against a trusted reference environment. Success is evidence—not vanity metrics.
Workload fit
- Divisible and container-compatible
- No identifiable patient data on donor nodes
- Reference outputs available for comparison
- Tolerant of variable contributor hardware
Measures
Propose a workload - Completed and validated compute hours
- Result accuracy against the reference run
- Completion rate and recovery from node loss
- Contributor retention at 30, 60, and 90 days
From prototype to proof
A disciplined path to the first research result.
- NowTechnical prototype
Donor, researcher, dispatch, and job-management surfaces.
- NextSecurity validation
External review, adversarial testing, and remediation.
- PilotInstitutional workload
Measured results against an approved reference environment.
- ScalePublic donor network
Open participation backed by published impact reporting.
The founding network
Four ways to move the project forward now.
Apoptosis is deliberately opening with partners, reviewers, and a controlled pilot—not a premature public download.
Research institutions
Bring the first reference workload.
A suitable pilot is divisible, container-compatible, benchmarkable against trusted infrastructure, and free of identifiable patient data on donor machines.
Scope a pilotSecurity reviewers
Challenge the architecture before donors trust it.
The next readiness gate is an independent review of isolation, authentication, task integrity, and the donor threat model.
Review the modelHardware and funding partners
Fund proof, not promises.
Support the security review, controlled pilot infrastructure, contributor hardware, or a mission-led launch to the gaming community.
Discuss sponsorshipFuture compute donors
Be first when the public agent is ready.
Join the launch network now. Installation will open only after independent security review and pilot remediation.
Join the launch listClear answers
What is ready—and what comes next.
Can I install the donor agent today?
Not publicly yet. A technical prototype exists, but installation opens only after independent security review, remediation, and a controlled institutional pilot.
What makes a research workload a good fit?
Strong candidates are divisible, container-compatible, tolerant of variable hardware, benchmarkable against a trusted reference, and do not require identifiable patient data on volunteer machines.
Is this replacing institutional HPC?
No. The goal is to complement trusted research infrastructure with elastic capacity for suitable workloads—not replace secure systems, ethics review, or institutional oversight.
Is Apoptosis already a registered charity?
The project documentation includes Canadian nonprofit, charitable registration, and US 501(c)(3) preparation materials. Those are filing plans, not a claim that current contributions are tax-deductible.
How can an organization help now?
Research teams can propose a pilot workload. Security experts can support architecture review. Hardware and funding partners can help bring the first independently validated pilot online.
Founding partners wanted
The hardware exists.Let's aim it at something that matters.
Bring a pilot workload, challenge the security model, or help fund the first trusted proof of volunteer compute for cancer research.
get-apoptosis.online