plumbobLLC
Early Stage
Plumbob LLC · Platform & Products in Active Development

Grounding AI in
deterministic execution.

AI products repeatedly rebuild the same dangerous distributed-systems machinery around model calls. Plumbob is building the execution platform to solve it.

The Plumbob Thesis: Products define what the work means. Plumbob owns the distributed execution required to fulfill it: immutable inputs, pinned model releases, durable Cloudflare Workflows, atomic spend reservations in micro-USD, provenance, and fenced erasure.

Explore the Thesis First Product: Cremaflow What We're Building Toward
PLATFORM ENGINECloudflare Workers + Workflows
STORAGE & AUDITD1 (State & Audit) + R2 (Content-Addressed)
FRONTIER AIClaude via Pinned WorkloadReleases
COMPANY STAGEEarly Stage · Pre-Commercial (Illinois, US)
Architectural Thesis

Products Own Meaning. Plumbob Owns Execution.

"Products own meaning. Plumbob owns reusable mechanism. Cloudflare owns infrastructure mechanics."

THE PLATFORM

Plumbob Cloud

Product-agnostic execution infrastructure that turns arbitrary AI interactions into bounded, durable, and auditable units of work.

  • Tenant-Bound Machine Identity: Scoped credentials isolate machine callers without cross-tenant data leakage.
  • Durable Cloudflare Workflows: Orchestrates multi-step jobs across interruptions with auditable terminal states.
  • Atomic Spend Governance: Monthly budget ceilings verified in D1 before inference dispatch.
  • Lineage-Aware Erasure: Purges input artifacts and all derived outputs through fenced cascades.
Inspect Platform Mechanics ↓
THE FIRST PRODUCT

Cremaflow

An in-development, local-first IoT application for specialty espresso extraction, physical sensor fusion, and deterministic replay.

  • Deterministic Local Core: High-frequency BLE sensor loop (scales, pressure, flow) operates 100% offline.
  • Frozen Evidence Bundles: Packages physical extraction telemetry into SHA-256 hashed immutable artifacts.
  • Claude Advisory Workloads: User-triggered shot analysis and recipe generation executed through Plumbob.
  • Deterministic Advisory Validation: Rejects or quarantines model claims that conflict with trusted sensor telemetry.
Explore Cremaflow Integration ↓
First Product Integration — In Development

Cremaflow: Grounding Claude in Physical Telemetry

Cremaflow is the first Product being built to consume Plumbob's execution platform, bringing nuanced reasoning to physical fluid dynamics without compromising local deterministic truth.

First Product Built on Plumbob · In Development

Cremaflow: Deterministic Telemetry & Grounded AI

Claude is Plumbob's first live frontier-model integration. Plumbob treats model inference as a bounded platform operation rather than an unconstrained API call: each workload pins its model and execution policy, reserves spend before dispatch, records observed model and usage provenance, validates structured output, and keeps probabilistic advice downstream of Product-owned deterministic truth.

Cremaflow is the first Product being built against this model. Its espresso telemetry, replay, analysis and live guidance remain deterministic and local. When the user explicitly requests AI analysis, Cremaflow freezes the relevant evidence, sends that immutable request through Plumbob, and treats Claude's result as provenance-tracked advisory output subject to deterministic Product validation.

Deterministic physical system→Immutable evidence→Bounded Claude execution→Deterministic validation→Advisory user experience
LOCAL RUNTIMEDocker · Python 3.14 · SQLite
HARDWARE LOOPESPHome BLE · 50 Hz Telemetry
AI WORKLOADSSession Review · Recipe Drafts (M7C)
ADVISORY SAFETYDeterministic Advisory Validation
DOSE18.0 g
YIELD36.2 g
TIME28.4 s
AVG PRESSURE8.4 bar
RATIO1:2.01
9 bar6 bar3 bar0InfusionFull PressureTaper & Cut
Pressure (bar)
Flow Rate (ml/s)
Yield (g)
Raw sensor observations acquired at 50 Hz locally via BLE. Replay and extraction engines execute deterministically on-device without cloud dependencies.
Architecture Under Development

What Plumbob Is Building Toward

More than model invocation: higher-order capabilities for multi-step AI systems, built above an immutable execution kernel.

Plumbob is being built so Products can submit intent rather than orchestrate infrastructure. The platform owns the distributed execution required to fulfill that work: immutable inputs, pinned model and workload releases, durable recovery, bounded spend, provenance, cancellation, and erasure.

Above that execution kernel, Plumbob is developing reusable higher-order capabilities for multi-step AI systems:

Idempotent Kernel

Execution Envelopes

A Product submits immutable input under one idempotency identity. Plumbob owns Artifact admission, workload execution, retries and recovery, provider attempts, spend limits, cancellation, and erasure underneath it.

Orchestration Graphs

Released Multi-Step Programs

Products can compose immutable Workloads, search operations, and evidence acquisition into bounded directed acyclic graphs (DAGs) without turning the model itself into an unrestricted runtime orchestrator.

Provenance Capture

Grounded Research & Evidence

Search discovers candidate sources; Plumbob retrieves exact source bytes itself, records cryptographic provenance, and distinguishes model claims from evidence that the platform actually observed.

Separation of Concerns

Independent Evaluation

A Product can use separately released Workloads to evaluate outputs produced by other Workloads, preserving both results and their provenance rather than asking the same model to self-certify.

Controlled Loops

Bounded Adaptive Execution

Future loops and repair paths have declared iteration, operation, time, and spend ceilings. A model may select among released choices; it cannot invent new runtime authority.

Consequential Actions

Controlled Effects

Longer term, Plumbob can turn consequential external actions into immutable commands with explicit authorization, idempotency, uncertain-outcome reconciliation, and optional human release.

Execution Kernel

How It Works: Cloudflare-Native Architecture

The distributed execution foundation that gives Products durable orchestration, spend governance, and lineage-aware erasure.

Identity & Isolation

01. Tenant Machine Ingress

Authenticated machine principals act strictly within stored Product and Tenant authority. Request bodies never include tenant identifiers.

  • Bearer credentials (plb1.<id>.<secret>) derive caller tenant identity and allowed scopes.
  • Per-source flood guards (1,000 req/60s) and per-principal fairness rate limits (120 req/60s).
  • Administrative control plane (/control/v2/*) strictly partitioned behind Cloudflare Access JWT verification.
HTTP / APIVerified Guarantee
// Ingress authorization & machine context
POST /v2/workload-runs HTTP/2
Host: api.plumbob.cloud
Authorization: Bearer plb1.cred_01j8m4q...secret
Idempotency-Key: idemp_98fa7c20e1
Content-Type: application/json

{
  "workload": "cremaflow.session-review",
  "inputArtifactId": "art_01j8m4z...",
  "expectedDigest": "6d2bc949298a09f48dd9ff99..."
}
System Guarantees

Non-Negotiable Engineering Guarantees

The guarantees enforced by database triggers, write leases, CI verification gates, and architectural linting across all Plumbob systems.

Atomic Spend Reservations

Worst-case cost reservations before provider dispatch

Every live execution atomically reserves estimated spend in micro-USD against Product and Tenant monthly ceilings before invocation. If a request cannot fit within budget, it fails immediately without reaching the provider. Retries have a strict limit of 0.

Exact-Bytes Authorization

Cryptographic binding to SHA-256 byte digests

User egress consent authorizes the exact immutable admitted body hash, selected privacy classes, and actor. Attempting to upload altered bytes under the same authorization triggers an immediate conflict error.

No Hallucinated Authority

Advisory outputs downstream of canonical truth

Probabilistic AI outputs are strictly advisory. A model can never mutate canonical domain state, edit SQLite journal tables, publish recipes automatically, or trigger physical machine actions without explicit user review.

Deterministic Foundation First

100% offline edge operability

Physical sensor acquisition, deterministic extraction algorithms, and live guidance run entirely local-first. Cloud infrastructure or external model provider outages never disrupt instrument operation or brewing.

Lineage-Aware Fenced Erasure

Fenced deletion markers across storage

Tenant content is retained only until erased. Calling DELETE /v2/artifacts/{id} triggers a cascading purge across all raw model outputs and derivative results in R2, leaving only immutable D1 audit metadata.

Model Pinning & Strict Schemas

JSON Schema 2020-12 enforcement

Execution policies pin exact provider model identifiers, gateway parameters, and strict JSON Schema definitions. Workloads reject silent prompt drift, non-conforming responses, and unverified model changes.

Company & Architectural Fit

Built for Frontier-Model Products

Plumbob LLC is an early-stage company established in 2025, based in Illinois, United States. We design and build execution infrastructure for AI-native software and edge systems.

Early-Stage Transparency: Plumbob Cloud and our initial products (including Cremaflow) are in active internal engineering. We are pre-commercial and pre-revenue, focused entirely on distributed-systems rigor, deterministic validation, and safety.

WHY CLAUDE MATTERS TO OUR ARCHITECTURE

Deploying Frontier Models Responsibly

Deploying frontier models like Claude in high-consequence applications requires moving from “respond to a prompt” into real software infrastructure:

  • Deterministic Evidence Boundaries: Claude receives immutable evidence bundles extracted from ground-truth local systems, never direct or unmetered database access.
  • Bounded Execution & Structured Outputs: Pinned releases, strict JSON Schema 2020-12 enforcement, and pre-execution monetary spend caps prevent runaway token consumption and prompt drift.
  • Downstream Advisory Validation: Model outputs are treated as probabilistic hypotheses, verified by product-owned deterministic rules before reaching users or physical actuators.

Get in Touch

Inquiries & Partnerships

We welcome technical discussions with systems engineers, researchers, and prospective partners interested in durable execution for AI systems.

APPROVED COMPANY ADDRESShello@plumbob.ai
LEGAL ENTITYPlumbob LLC
JURISDICTIONIllinois, United States
PRIMARY DOMAINplumbob.ai