AI-assisted kitchen inventory and recipe planning that turns what you already have into meals.
AI decisions should be traceable before they become consequential.
InTellMe builds evidence-governed research, verification, and decision infrastructure for organizations operating in high-uncertainty environments.
Core technology
The work that gives the house its name.
TruVector
Validation and pilot developmentPre-execution arbitration of AI assertions before downstream systems act.
- Allow
- Review
- Block
TruVector sits in front of consequential action. It weighs evidence, independence, and disagreement, then issues a decision a downstream system can obey.
ResearchOne
Private beta — early access availableStructured deep research that keeps sources, contradictions, reasoning lineage, and report contracts intact.
- Sources stay attached.
- Contradictions are preserved.
- Reasoning lineage is kept.
- Reports obey a contract.
What is built, and what is not.
Each capability below carries its own state. Nothing is described as finished because the architecture calls for it, and the two unproven claims are named as unproven rather than buried.
Evidence object and decision path The structured result a downstream system reads, and the Allow · Review · Block boundary it obeys. Schema and boundary are fixed and running. Implemented
Contradiction and divergence Detecting that sources disagree, and carrying that disagreement into the decision instead of averaging it away. Detectable and measured; calibration is not settled. Under validation
Independence-aware evidence Discounting agreement between sources that share an origin, so that one claim repeated five times does not score as five confirmations. Under validation
Quorum The rule that decides how much independent agreement a claim needs before it may pass, given what acting on it would cost. Specified, not built. Planned
AHE / HRA Adjudication under conflicting evidence where no source is authoritative. Stated as a claim with a stated way to be wrong, and not yet demonstrated. Hypothesis
Kinematic validation Treating a claim's movement through evidence over time as a signal in its own right. The most speculative element here, and the reason the research is research. Hypothesis
Implemented — runs, and measured
Under validation — protocol written, results partial
Planned — specified, not built
Hypothesis — a claim with a stated way to be wrong
Where this sits against published work.
Pre-execution gating for AI systems is an active area, and there is published work describing an Allow / Deny / Escalate decision in front of an agent's action. Two of those decisions look alike from the outside.
They are not the same decision. The published approaches ask a permission question: is this actor allowed to take this action, under this policy? The answer is a lookup, and it is right or wrong.
TruVector asks an evidentiary question: given conflicting sources of unequal independence, how well supported is the assertion this action rests on — and is that support enough for what the action would cost? The answer is a judgement under uncertainty, and it carries its own error bar.
Policy gating and evidential adjudication compose. A system can want both, and neither substitutes for the other.
Commercial ventures
Products the infrastructure made possible.
Instant technical SEO audit of a live page.
AI-assisted apparel design, with Printful-integrated placement, mockup, checkout, and fulfillment. The storefront is open on live payments; it has not been marketed and has not taken a first order.
Research
A public observatory.
wAether
Live research productA live observatory of solar, galactic, and terrestrial telemetry, built on official government data sources — named because the live site names them.
- NOAA SWPC
- USGS
- NASA DONKI
- NOAA OVATION
Pricing starts at $4.99/month or $49.99/year.
Company
Qualification, applied to machines.
InTellMe is built and run by Michael Brandon Lane, from Johnson City, Tennessee.
He trained as a chemical engineer at Virginia Tech and took a master’s in chemistry at UNC Greensboro, then worked in industrial materials — including the lithium-battery separator programme where he is a named inventor on Celgard’s patent family for coated separators, WO 2019/143483.
That work is why this company exists. A separator is a safety part: when a claim about one is wrong, the failure is a fire, not a bug report. Nothing ships on a plausible result. It ships once you can say how it was measured, across what range, within what error, and what observation would have shown you were wrong.
Most AI systems have no equivalent step. They produce an assertion and act on it. TruVector is that missing step — the discipline of qualification, applied to a machine’s claims before anything downstream obeys them.
Current activity
Validation is labeled as validation.
Pilots are labeled as pilots. A public page is not a completed experiment. Work in progress is listed as work in progress.
- TruVector — validation and pilot development.
- NSF SBIR Project Pitch — in preparation. Pursued, not awarded.
- ResearchOne — private beta.
- SAVR — web beta and Android testing.
- Golden Goose Tees Studio — storefront in beta; unmarketed, first order still ahead.
- wAether — live.
Where the work lives
Every property, and its address.
Nothing here is a mockup. Each of these is a running site you can open now, at the stage it is actually in.
- truvector.science TruVector Pre-execution arbitration of AI assertions. Validation and pilot development.
- waether.space wAether Observatory of solar, galactic and terrestrial telemetry. Live, paid.
- savr.cam SAVR Kitchen inventory and recipe planning. Web beta, Android in testing.
- goldengoosetools.com Golden Goose Tools Instant technical SEO audit of a live page. Live.
- goldengoosetees.com Golden Goose Tees Studio AI-assisted apparel design and fulfillment. Storefront in beta.
- researchone.io ResearchOne Evidence-governed research workspace. Private beta.
Funding
Two tracks, kept unmixed.
Track A — TruVector research funding
An NSF SBIR Project Pitch is in preparation. It is being pursued, not awarded. The commercial applications are not the SBIR work.
Track B — Company investment
Pre-seed and strategic capital. Use of funds, by category:
- TruVector validation and pilot operations
- ResearchOne private-beta hardening
- SAVR launch: evidence, evaluation, and documentation
- Company operations
Contact
Reach the people doing the work.
InTellMe
326 Delaware StreetJohnson City, TN 37604
United States
- Telephone(747) 322-1977
- Generalinfo@intellmeai.com