Rillmark An institution for physical practice
Pre-founding · seeking partners
§ 00 · Thesis

A new way to think about physical practice.

Rillmark is an infrastructure company in formation, building an open standard for acts with typed accountability: production, research, supply chain, regulated work, and the rest of the surface where typed events with provenance carry institutional weight. It begins where most institutions begin: with the things makers and small workshops already do every day. The specification is open and will be Foundation-stewarded; a for-profit company will build the working substrate. Both are structured for a fifty-year horizon.

Rill, n. — a small stream carved by patient water. Mark, v. — to notice, to record, to leave a trace.   Small acts, repeated, that cut a channel the next generation can follow.
§ 01 · Why this, why now

Most of the work the world depends on is recorded after the fact, if it is recorded at all. A leatherworker's sense of which thread pairs with which needle lives in muscle and margins. A pharmaceutical batch streams seventy-two hours of sensor readings into a system the QC lab next door cannot read. A second laboratory reproduces a published protocol, gets different results, and has no way to say what specifically differed. A heritage carpenter, the last practitioner of a regional joint, retires without a successor.

Fragments are captured by ERP, PLM, MES, LIMS, electronic lab notebooks, recipe apps, registries. Every one of them is a record format: a way of writing down what already happened, in a schema the next system cannot read. The work runs in one place and is described in another, and the two drift apart from the moment the work ends. What survives cannot be searched, cited, audited, or replayed across domains.

Rillmark's thesis is that these are one problem refracted through different magnifications, and that the fix is structural. Six primitives — Tools, Materials, Techniques, Workflows, Projects, Events — describe acts rather than write-ups of acts: a project is a recipe, an event is its execution, and the run is the record. Doing and documenting stop being two activities that drift; they become one. Named consistently across craft, industry, research, regulated work, and emergency response, those records compose into a commons worth more than any single application built on top of it.

Principle 01

Your data, your format.

Every record the software touches is exportable without our cooperation: portable canonical JSON, content-addressed, verifiable. Records are plain JSON in a Git folder you own, readable without us. The ownership guarantee is in how it's built.

Principle 02

A commons held in trust.

The Open Practice Graph Language specification and the shared knowledge layer will be stewarded by the Rillmark Commons Foundation, a separate nonprofit entity structured so it can fork the for-profit if the company misbehaves. This is a structural protection, not a promise.

Principle 03

Fifty-year horizon.

The point isn't to raise and exit. The point is to still be around when your apprentice's apprentice opens your notebook.

§ 02 · The First Product

A free, local PM tool for makers, and the vehicle for the commons.

Shipping · local-first

HideSync

Reference implementation · free

HideSync is the working reference implementation of OPG-L 0.6, a free, local tool that records the full lineage of how you make things. Every record is a JSON manifest in a Git-tracked folder you own; nothing is held hostage to a server or a subscription. The hosted Rillmark services come later: a SaaS for small workshops and SMEs first, then enterprise, regulated industry, and research. All of it builds on the same standard; your HideSync records carry forward without lock-in.

The six primitives

The Open Practice Graph Language (OPG-L) specifies six primitives that capture the structural content of physical practice across any domain. The hard part of documenting a craft isn't writing things down. It's naming the things worth writing about. Rillmark picks six, just six, and uses them consistently. Small enough to keep in your head; general enough to fit leatherwork, laboratory protocols, aerospace manufacturing, and fermentation.

01 — Tool

Tool

A thing you use that does not get consumed. An awl, a cast-iron pan, a #7 jointer plane, a fermentation vessel.

example: "round awl · Osborne #477 · sharpened 2025-12"
02 — Material

Material

A thing you use up. Leather, wood, flour, wax, pigment, thread, solder. Tracked by origin and lot when that matters.

example: "Wickett & Craig · 8oz · nat. russet · hide #A-104"
03 — Technique

Technique

A repeatable way of doing something. Saddle stitch, pâte brisée, dovetail layout, brine cure. The unit of learned skill.

example: "saddle stitch · 6 spi · angled awl"
04 — Workflow

Workflow

An ordered sequence of techniques applied to materials using tools. A recipe, a cut list, a build order.

example: "card-holder, v3 · 14 steps · ~3.5h"
05 — Project

Project

What you set out to make, and the plan for making it. The bag for your sister, Sunday's loaf, the chair you're halfway through.

example: "commission · M. Reyes belt · due 2026-05"
06 — Event

Event

A dated record of work actually done. Sharpening. Receiving a hide. Finishing a project. The doing that ties history together.

example: "2026-02-14 · stropped awl · felt pull again"
01
Tool
A thing you use that does not get consumed. An awl, a cast-iron pan, a #7 jointer plane.
02
Material
A thing you use up. Leather, wood, flour, wax, pigment, thread.
03
Technique
A repeatable way of doing something. Saddle stitch, pâte brisée, dovetail layout, brine cure.
04
Workflow
An ordered sequence of techniques applied to materials using tools. A recipe, a cut list, a build order.
05
Project
What you set out to make, and the plan for making it. The bag for your sister, Sunday's loaf, the chair you're halfway through.
06
Event
A dated record of work actually done. Sharpening. Receiving a hide. Finishing a project.
§ 03 · The commons

A shared layer of knowledge that no one owns.

Open specification · Foundation-stewarded

The Rillmark Commons Foundation will steward not a database of shared objects but a specification wiki, comparable to Wikipedia for physical practice. Three categories, all content-addressed, hash-pinnable, and citable:

Canonical references — primitives grounded in external standards. Stable parents against which practitioner diffs inherit. Community specifications — techniques and methods with no external standard, defined by contribution and consensus over time. Most craft-domain techniques live here. Exemplar projects — complete, buildable primitives contributed by users as references. Full lineage, shared for others to fork or cite.

The knowledge of how to make things belongs to the people who make things.

A working proto-commons is already live: a public GitHub repository of content-addressed primitives, category indexes, and citeable bundles, CC-BY-4.0, accruing since launch. It is the seed the Foundation inherits. Browse the proto-commons →

Commons publication is opt-in per record. A user can hold every record at private indefinitely and use Rillmark as pure PM/ERP with zero commons exposure; the visibility gradient is a property of the standard, not a policy of any one platform.

When the Foundation is established, the specification stewardship transfers from Rillmark to the Foundation by design: OPG-L under CC-BY-4.0, reference implementations under Apache-2.0. Data expressed in OPG-L is portable to any conformant implementation that exists or will exist. The two-layer structure that makes this load-bearing →

§ 04 · Pricing model

HideSync is free and stays free. The hosted Rillmark SaaS, a separate and later product, would have tiers: Free — three workshops, full data entry, export; Power — cost tracking, analytics, unlimited workshops; Professional — e-commerce, sales, supplier tools; Compliance — pre-built DPP and regulatory submission workflows; Enterprise — teams, multi-location, API, custom integrations. Upgrade reveals insight over data that was always captured; downgrade is not data loss. Pricing is not finalised; this page will say so plainly when it is.

§ 05 · A door
What follows is the institutional half of Rillmark, for a CTO, a venture-studio partner, early makers, and anyone who reads structural thinking for pleasure. It is denser than the surface. It is also the point.
Rillmark
Deep dive · prepared for partners · read ~12 min
↑ back to surface
For partners · CTO · venture studio · early makers

A fifty-year institution, not a four-year exit.

Rillmark is a Standard/Infrastructure play with two layers. Layer 1 is OPG-L, an open standard for acts with typed accountability, stewarded under the Rillmark Commons Foundation. Layer 2 is the OPG-L Node: organization-operated implementations of that standard. Same shape as Linux Foundation + Red Hat, ICANN + Verisign + country-code TLDs, CNCF + Kubernetes implementers, W3C + browser vendors. Not founded yet, deliberately: the structure can be the thing rather than something retrofitted once incentives are already pulling the wrong way.

We are looking for three kinds of partner: a technical co-founder (CTO) to own Rillmark Inc.'s canonical commercial Node, the Headless OPG-L backend and the first commercial heads; a venture-studio partner literate in two-layer Standard/Infrastructure plays, patient with the standard's decade horizon while the company moves at startup speed; and early makers willing to use HideSync today and seed the proto-commons. A Chief Ecosystem Officer hires in at seed to lead the Foundation Institutional Relations Subdivision.

§ I · Architecture

Built to outlast the company that ships it.

OPG-L is computational, not merely descriptive: a project is a program, an event is its execution: the doing and the record of the doing are one artifact. That is what separates it from the provenance formats it interoperates with, which only serialize what already happened.

HideSync is the reference implementation: a locally-hosted, three-service realization of OPG-L 0.6 in production-grade software: content-addressed storage, immutable lineage, amendment-Event corrections, provable upstream commitments at visibility boundaries. The same model serves a hobbyist's wallet and a pharmaceutical batch record from one spine: the standard's, not a shared codebase. Rillmark's commercial backend will be a separate, ground-up build; the bridge is the wire format, so records made in HideSync export into it without lock-in.

The architecture is CQRS-centric and content-addressing-centric. Every projection (graph, analytics, search, the commons plane) is a consumer of the same write log. This is what makes the commons plane physically separable onto Foundation-owned infrastructure: a deployment change, not a data-model migration.

This inverts the usual SaaS assumption. The product is a view over an open standard. The strategic consequence: Rillmark's durability depends on being useful, not on being sticky. The commons is the asset: not the code, but the accumulated attributed knowledge and the governance structure that keeps it neutral.

Bridging-as-business-model. A bridging standard is credible only when bridges exist that use it. Rillmark's own multi-head architecture is the running demonstration. Every head is a production proof point for OPG-L: SaaS PM/ERP for SMEs, an Alias & Licensing Service for cross-jurisdictional provenance, a Gated Academia Portal for citation-grade reproducibility, an Audit & Validation layer for regulatory neutrality, a Search & Rescue head for multi-agency emergency coordination. Competitors selling an alternative standard would have to build all of these themselves to match the proof. Defensibility is structural, not feature-based.

§ II · Three operator categories

One standard. Sovereign implementations.

The OPG-L Node — not Rillmark Inc., not the Foundation — is the federation primitive. Any organization-operated implementation of OPG-L is a Node. Nodes emit OPG-L records, consume OPG-L records, and federate via the standard. Three operator categories carry the load:

Operator 01 · Canonical Commercial Node

Rillmark, Inc.

For-profit · builds the engine and the heads · operates the commercial Node

Develops and builds the Headless OPG-L backend and the heads on it, both its own and the infrastructure the Foundation operates, working with domain partners where a head requires them. Operates the commercial heads: the SaaS PM/ERP for SMEs and the commercial Research Planner. A public-interest head can run under Rillmark Inc. before the Foundation is ready to carry it, under explicit Foundation license, with migration committed. Pays the Foundation tiered OPG-L licensing fees, the structural mechanism that funds the Foundation independent of grant cycles. Analog: Red Hat in the Linux ecosystem.

  • Headless OPG-L backend + SaaS PM/ERP (first head)
  • Free, Power, Professional, Compliance, Enterprise tiers
  • Builds the head infrastructure the Foundation runs onward
  • Records are portable JSON in Git, no proprietary export tool
  • Not a gatekeeper of the specification or the commons
Operator 02 · Foundation-Operated Nodes

Rillmark Commons Foundation

Nonprofit · operates public-interest heads · stewards OPG-L

Operates heads where commercial governance would compromise the mission. The Alias & Licensing Service for cross-jurisdictional provenance is the canonical case; the Gated Academia Portal, Audit & Validation, Search & Rescue, the Commons, and the academic Research Planner follow as each stands up. The Foundation does not build this infrastructure: Rillmark Inc. builds it, and the Foundation runs it onward as public-interest service. eIDAS-class faithful-recording posture. Stewards the OPG-L specification and conformance program.

  • Stewards OPG-L under CC-BY-4.0 + reference implementations under Apache-2.0
  • Signed nightly Merkle manifests, commons integrity independently verifiable
  • Can, by charter, fork if the company misbehaves
  • Funded by OPG-L licensing fees, directed contributions, grants, MoUs

Operator 03 · Independent Sovereign Nodes. Operated by entities that are neither Rillmark Inc. nor the Foundation: First Nations communities, academic institutions, regulators, NGOs, sectoral consortia. Each independent Node is governed by its operator on its own terms. The Foundation facilitates independents (technical scaffolding, legal templates, federation-protocol coordination) but never owns or operates them. Sovereignty is preserved by construction, not by policy. Analog: country-code TLDs in DNS.

Allocation follows function, and phase. A head lives with the Foundation when its operator-shape would otherwise compromise it: neutral registry, academic legitimacy, regulatory neutrality, multi-agency trust. A head with clean commercial economics lives with Rillmark Inc. And a head can run under one operator on its way to the other: the Alias & Licensing Service can launch under Rillmark Inc. on explicit Foundation license and migrate at activation. Two things do not move: Rillmark Inc. builds the infrastructure, including the Foundation's; the Foundation provides the standard, including to Rillmark Inc.

Operates vs. facilitates is the legal hinge. The Foundation operates head infrastructure under eIDAS-class posture. The Foundation facilitates the substantive work that uses those heads (research projects, SAR operations, indigenous-data work) without carrying operational liability for it. Conflating the two exposes the Foundation to liability it cannot insure against.

The fork clause is not ornamental. Once the Foundation holds the specification and the data, and because the application is only a view over that data, any sufficiently motivated group of practitioners can reconstitute Rillmark around a new client without losing the accumulated commons. This is what "no data hostage" means structurally. Today the export, the running software, and the public commons are already checkable; the specification, the Foundation charter, and the licensing schedule become checkable as the Foundation forms. By design, everything Rillmark commits to is built to be independently verifiable.

The funding flows in a circle that keeps both sides honest. Rillmark Inc. licenses OPG-L from the Foundation, the same shape as Red Hat's dues to the Linux Foundation, and a commercial operator paying for the standard is the clearest proof of its value. The capital pools stay separate: equity investors cannot push the Foundation to commercialize, and grant funders cannot push the company off-commercial.

§ III · Bridging-as-business-model

Four doors. One federation.

Defensibility is the running demonstration. A bridging standard is credible only when bridges exist that use it. Rillmark will address four distinct surfaces with the same primitives, the same specification, and the same Headless backend, entered in deliberate sequence. Compliance falls out as a by-product, not a separate product line.

The deeper reason for the portfolio: standards win slowly, and each head shortens the curve. Every head onboards a specific cohort straight into the federation: SMEs, customs and insurers, researchers, regulators, emergency agencies. Commercial traction and standard adoption become the same motion seen from two layers. Two clocks run at once, the commercial layer on startup cycles and the standard on the category horizon, and they compound in one direction: every subscription is an emitter joining the federation. One head is a proof point; the whole portfolio is what makes the long horizon tractable, not aspirational.

Door 01 · B2C

The long tail of makers.

Solopreneurs, artisans, small workshops, community practitioners

Cheap, near-frictionless. Documents practice the way makers already think about it; compliance and provenance fall out as a byproduct. Populates the commons before the first enterprise customer arrives. Viral by artifact: every shared protocol, every exported record carries the standard into a new community.

  • Free + Power tiers; prosumer SaaS pricing
  • Builds the corpus that makes the commons valuable
  • Distribution: GitHub, craft forums, makers' word of mouth
Door 02 · B2B + EU regulatory

Enterprise, regulated industry, customs.

Pharma, manufacturing, supply chain, luxury, CITES, DPP/ESPR

Same spine, different head. CQRS and content-addressed lineage are exactly what regulated industries already pay millions for in bespoke ELN, MES, batch-record, and customs systems. OPG-L is architecturally what Digital Product Passport is trying to mandate into existence: provenance-bearing, content-addressed, cross-domain. Every new EU product regulation expands the addressable market with little new code. Procurement cycles are long; the commons existing first is what makes the conversation possible.

  • Enterprise tier + Alias & Licensing head for cross-jurisdictional provenance
  • Compliance is a configuration, not a rebuild
  • Pays the highest ARPU; never builds the asset
Door 03 · Academia + heritage

Research, citation, sovereign archives.

Universities, labs, indigenous archives, heritage institutions

Citable hashes; exemplar projects published as primary references. Researchers cite an OPG-L hash the way they cite a DOI, a credibility flywheel for the standard and free distribution into every adjacent discipline. The Foundation's natural constituency. Methods / results / inferential reproducibility fall out of OPG-L's lineage discipline.

  • Free for academic use; CC-BY-4.0 at the Foundation
  • Exemplar projects: complete, citable, fork-able
  • Foundation-stewarded; channel for directed contributions
Door 04 · Sovereign Nodes

Independent operators, federated.

First Nations communities, regulators, NGOs, sectoral consortia

Sovereign Nodes are the federation primitive for operators whose data sovereignty must be preserved by construction: operated by the sovereign entity on its own infrastructure under its own governance, OCAP / CARE-aligned where indigenous data is involved. The pattern supports an interoperability layer between a sovereign operator's data and academia (and, where the operator chooses, commerce and regulation) without ceding control to either side. Federation is permissionless at the standards level. The Foundation facilitates; it does not own, and pre-seed that means conversations, not committed examples.

  • Operated by sovereign entities, not by Rillmark or the Foundation
  • Visibility-boundary mechanism preserves sovereignty cryptographically
  • cc-TLD-in-DNS is the analog

The sequence matters. The long tail does the infrastructure work that makes the premium end of the market possible. The enterprise customer does not build the commons; the solopreneur does. By the time a brand wants to make verifiable claims about its supply chain, the records are already there, contributed by the producers they buy from. That is the asset.

§ IV · Roles sought

Three shapes of partner. Wanted now.

Role 01

CTO · technical co-founder

co-founder equity · full-time

Deep experience with distributed systems, content-addressed storage, and CQRS architectures. Standards instinct, not just product-shipping instinct. Comfortable owning a reference implementation of a public specification and a conformance program alongside it. Reference profiles: shipped CNCF / Apache / W3C work and held senior engineering roles at infrastructure companies (MongoDB, Confluent, HashiCorp, Cloudflare, Stripe at equivalent stage). Patient-capital temperament. Wants true ownership and a partnership, not an employee role.

Role 02

Venture-studio partner

first-institutional cheque · studio structure welcome

Two-layer Standard/Infrastructure-play literate. Has backed (or studied) Linux Foundation companies, CNCF projects, W3C-track plays, open-source-commercial companies. Understands the two clocks: the SaaS lets Rillmark Inc. move at startup speed (commercial cycles, revenue, proof points every six months) while Inc. and Foundation pursue the standard in parallel on the seven-to-ten-year category horizon. Patience is for the category, not the company. Network access for CTO and CEcoO recruiting, corporate introductions, academic partnerships, and indigenous-organization bridges. Operational infrastructure for dual-entity incorporation.

Role 03

CEcoO · Chief Ecosystem Officer

co-founder-level partnership · for-profit + Foundation

PhD anthropologist or adjacent. Cultural-translation skills across academic, regulatory, corporate, and indigenous governance contexts. At seed, leads the Foundation Institutional Relations Subdivision: MoU pipelines, directed contributions to OPG-L-related Foundation programs, government partnership relationships, academic memberships, indigenous-facilitation relationships. Originates commercial BD that benefits Rillmark Inc. in parallel. Hires in at seed, once the for-profit is formed; the conversation starts now.

Early makers are also welcome. If you practise a craft and want to seed the commons and document your techniques in OPG-L, write to [email protected]. Founding-contributor credit will carry into the Foundation when it forms.

§ V · The long thesis

Why Rillmark outlasts the people who start it.

  1. Physical practice is undocumented in ways software work isn't.

    There is no GitHub for hands. The knowledge of how to make things is transmitted person-to-person, notebook-to-notebook, and lost generationally. A typed, content-addressed substrate is the minimum viable intervention.

  2. The closure-at-six kind universe is the moat.

    Six primitives — Tool, Material, Technique, Workflow, Project, Event — cover the structural roles of acts. The strongest candidate for a seventh (commitment / obligation, drawn from REA-2) does not survive scrutiny; planned-events plus cascade-close plus amendment recover it through structure rather than ontology expansion. TCP/IP is small. HTTP is small. SQL's core is small. Each bridges enormous heterogeneity by refusing to grow.

  3. A commons is only a commons if it is structurally ungovernable by one party.

    The two-layer structure isn't cosmetic. Federation is permissionless at the standards level: the Foundation cannot revoke a Node's federation status; Rillmark Inc. cannot refuse to interoperate with a competing implementation. cc-TLD-in-DNS is the analog. The commons remains neutral in twenty years when the market has moved and the original team has not, because the architecture is what protects it — not the founders' good intentions.

  4. Starting with a real product is the honest sequence.

    Most institutions that want to serve a community start with the institution and work outward. Rillmark starts with a shipping PM/ERP product — in the hands of real makers — and earns the right to build the institution around what the software reveals. HideSync — a local, working reference implementation of OPG-L 0.6, built solo across three services — is the appreciating asset that lets a technically-trained founder enter the venture-studio conversation with proof rather than slides, and makes the CTO a deliberate first hire rather than a gap to fill. A standards body cannot ship a usable product. A startup selling SaaS cannot credibly promise ten-year lineage integrity. Rillmark is the structure that does both.

  5. Durability compounds.

    Every notebook preserved is a line of inheritance re-established. The return on infrastructure like this is not linear — it is the kind of return you only see after a generation, which is exactly why almost no one builds it.

If any of the above fits you, write.

Please include a paragraph on why you think this is interesting, and another on what you'd want to build. We read every message. We answer more slowly than we'd like and more thoughtfully than most.

For CTOs & CEcoOs
Write to the founder directly.
For venture-studio partners
Expect a longer conversation.
For early makers
Seed the commons with us.