Skip to content
Install切换到中文: EN/中

support the lab

Support WhiteMagic Labs.

Share what you learned, work with us, or leave a voluntary tip. The local core stays free to use, with no metering or contribution requirement.

Free ways to support

  • ·Cite the project where it was useful.
  • ·Star the repository.
  • ·Tell one person who runs agents why local-first memory matters.

Tried WhiteMagic? Share a result, a question, or something that did not work. A sentence is enough. Tell us what happened. Setup help lives on the Install guide.

Ways to work together

Start with the question you want to answer. Enterprise covers scoped deployments, integrations, and evaluations.

Research collaboration

Reproduce or extend an evaluation of agent memory and continuity. Bring a system or a question, and agree on a method and a public report together.

Integration partnership

Making WhiteMagic work well in your MCP client, framework, or deployment. Public write-ups either way.

Sponsored research

Fund a specific question with a public report at the end. Method stays the lab's; results stay published, losses included.

Grants and foundations

Non-dilutive support for the trust and verifiability work. No equity, no board seat, no licensing strings; funding never buys a preferred result.

Research support does not buy favorable results or change the method to suit a sponsor. Reports include failures as well as successes.

Public Infrastructure · Scoped Grant Work Packages

Fundable Open-Source Public Good Milestones

We welcome non-dilutive grant support from public interest technology bodies (such as the Sovereign Tech Fund and NLnet Foundation). Grants are scoped around modular technical deliverables:

Systems Security · Milestone A$30k – $60k

Cross-Platform OS Sandboxing & Landlock LSM Policy Profiles

Formalize Linux Landlock LSM kernel restrictions into portable policy profiles for developer toolchains, exploring macOS sandbox equivalents.

Open Standards & Proofs · Milestone B$25k – $50k

Standalone CPR Verifier & IETF SCITT / AER Test Vectors v0.6

Deliver standalone Rust/Python/WASM Causal Provenance Receipt (CPR) reference verifiers with conformance vectors for zero-network audits.

Public Evaluation · Milestone C$40k – $80k

Zero-Telemetry Agent Continuity Benchmark & Evaluation Harness

Build a reproducible, open-source benchmark suite measuring multi-session agent context recall and efficiency strictly without cloud telemetry.

Systems Security Engineering Posture: WhiteMagic focuses on deterministic OS kernel sandboxing, cryptographic provenance, and zero-egress resilience. We build open digital public goods and do not participate in speculative alignment tracks or involuntary restraint initiatives.

Support goes, in this order

  1. Compute – The machines that build, test, and evaluate the software.
  2. Hardware – Evaluation and reproducibility boxes, bought only after measuring.
  3. Infrastructure – Hosting for optional public services as traffic earns it.
  4. Research time – Benchmarks, evaluations, and the writing that makes results checkable.
  5. External audits – Paying people who do not work for us to try to break what we build.
  6. Public artifacts – Synthetic corpora, receipts, and reproduction packages.

The local core is never metered. Funding does not determine evaluation methods or which results are published.