// the platform
The Mad Botter is operational data infrastructure for the unconnected edge — deterministic capture, a sovereign data layer, and pipelines, deployed on your hardware and operated by forward-deployed engineers.
// capture, powered by vorpal
// layer 01 · deterministic process record
Most operational records are reconstructed: clipboards, end-of-shift data entry, an operator's memory of which bin went on which pallet. Audits, recalls, and disputes are all arguments about what happened — and reconstructed truth loses those arguments. The Capture layer records the floor from the machines, in order, as it happens. It ends the argument. Your compliance record shouldn't depend on what the operator remembers writing down.
x end-of-shift data entry, from memory
x clipboard → spreadsheet → argument
+ recorded from the machines, in order
// illustrative — not customer data
Hardware-agnostic ingestion — gateways, scales, label printers, barcode scanners, kiosks, and the serial-line equipment nobody else will talk to.
Offline-first with ultra-low-bandwidth sync. An ordered event log on your hardware — nothing lost when the link drops, nothing reordered when it returns.
Replay any window. Full audit trail. Compliance outputs generated as byproducts of the record — MIL-STD-129 labels, lot traceability, inspection records.
// layer 02 · sovereign data layer
PostgreSQL, Parquet, and DuckDB behind your boundary — not a vendor's cloud. Air-gap is a supported configuration, not a workaround: the record is written locally and syncs when the link exists. Glass box, open formats — the storage outlives any vendor, including us.
Commodity edge hardware — BeagleBone and Raspberry Pi–class gateways — feeding a data layer on your servers. Self-hosted over WireGuard where a link exists; fully inside the enclave where one doesn't.
The layer's lineage is open source — read the hardware abstractions and the lakehouse engine yourself instead of taking a write-up on faith.
// pipelines, powered by alice
// layer 03 · unification + BI
Before you can automate the edge, you have to untangle the past. Pipelines connects twenty-year-old, undocumented on-prem schemas to the core without breaking compliance — you inspect the generated SQL and a real-data preview, approve explicitly, and nothing executes until you have. Two pilots, one layer:
// self-serve pilot
One source, one Power BI dataset, a scheduled refresh — live before your next reporting meeting. Connect, inspect what the platform proposes, approve, done. No sales call.
// enterprise pilot
On-prem or single-tenant deployment for the estate the self-serve tier can't touch: custom sources, security review, procurement support, and schema unification that preps the ground for Capture.
Connect read-only → the platform proposes the transformation → you read the generated SQL and a real-data preview → you approve explicitly → scheduled, verified refreshes. Row data never reaches the model — only your instruction plus table and column names and types. Full data-flow answers live on the security page.
> review transform #142
SELECT jh_no AS job_id, jh_cst AS customer
FROM legacy.job_hdr;
preview: 200 rows · types verified
+ approved by you · scheduled nightly
// illustrative — not customer data
// deployments
Specifics rather than capabilities, because you can check specifics. All of this is in production — for a DC-area aerospace parts broker and MRO shop, a Cincinnati-area deep foundations contractor, and a large industrial agriculture operation.
Field gateways streaming environment data from high-value inventory; a rule engine grades breaches and alerts to email and Slack.
// capture + corePrint-ready PDFs and ZPL to Zebra printers; RFID records cross-reference job, label, shipment TCN, and the DoD scan.
// captureHand-typed reference numbers stitched into one relational job with a single identity from quote to delivery.
// pipelinesAsset identity and QR tags across five operating companies; cert history for an OSHA inspector in one scan.
// capture + core500+ endpoints, 70+ models, AWETA lines, bin fillers, scales, scanners. Self-hosted over WireGuard — ours to run.
// capture + core + pipelinesLot traceability from harvest bin to customer; cold chain verified at close; certification records as a byproduct.
// capture + pipelines// ownership · what you buy
You own your data, your infrastructure, your configuration, and your deployment artifacts. We retain the reusable platform core and license it to you; updates and managed operations run on a platform agreement.
Your operational data, the infrastructure it runs on, your configuration, your integrations, and the deployment artifacts — exports, schemas, runbooks, and credentials. All of it stays yours whether or not we keep working together, and none of it is hostage to a renewal.
The platform core — the capture engine, the data layer, and the pipeline runtime — is our intellectual property, licensed to you rather than sold. That is what makes a deployment take months instead of years, and it is why a pilot can be priced flat and low instead of quoted as a bespoke build.
Continued platform updates, security patches, and operations run against an agreed service level. Let the agreement lapse and the deployment keeps running on your hardware with your data — you simply stop receiving new platform versions and we stop carrying the incident response.
// compliance
Engineered for ITAR, EAR, CMMC, and NIST environments. Air-gap deployment is a supported configuration, not a workaround — the platform runs on your hardware, behind your boundary, and syncs when the link exists. We are designed to support your compliance posture; we do not certify compliance. And we'll say exactly that to your auditor.
// the front door
$1,500 flat. Thirty days, one production line, verified by replay at day thirty — you judge the platform on the record it produces. The full $1,500 credits against your first platform invoice if you continue. The second door: the self-serve pipelines pilot, $149/mo with a 14-day trial, live before your next reporting meeting.
thirty days · one production line · verified by replay
// faq
No. Capture doesn't guess what will happen; it records what did. Alerting on sensor thresholds is a feature; the product is the record.
Yes. Offline-first by design — the record is written locally and syncs when the link returns. Air-gapped deployment is a supported configuration.
Commodity edge hardware — BeagleBone and Raspberry Pi–class gateways — and your servers, not our cloud.
Your instruction plus table and column names and types. Row data does not reach the model. You read the generated SQL and a real-data preview before anything executes, and nothing runs until you approve it explicitly.
You own your operational data, the infrastructure it runs on, your configuration and integrations, and the deployment artifacts — exports, schemas, runbooks, credentials. We retain the reusable platform core (the capture engine, the data layer, the pipeline runtime) and license it to you. That split is what makes a deployment take months instead of years, and it is why a pilot costs $1,500 rather than the price of a bespoke build.
The deployment keeps running on your hardware with your data, in open formats, and you keep every artifact. What stops is new platform versions, security patches, and managed operations — those run on a platform agreement. There is no kill switch and nothing phones home to disable itself.
The First Line Pilot is $1,500 flat — thirty days, one production line, machine-verifiable finish. The self-serve pipelines pilot is $149/mo with a 14-day trial. Full platform deployment is scoped against your actual systems.