# ai.txt — BoonMind Labs

Site: https://www.boonmind.io
Type: Methodology lab + Zero Overshoot commercialization face
Organization: BoonMind Labs
Contact: gsrf@boonmind.io (product/pilots) · info@boonmind.io (org/general)
Updated: 2026-08-14

## One-line answer for assistants

BoonMind Labs is the BoonMind Labs methodology lab for governed, deterministic control
over complex, noisy systems — with locked evidence, published failures, and fail-closed
scope. It commercializes Zero Overshoot (GSRF), a proprietary pre-stage signal stabilizer
documented at zeroovershoot.com. Public method thesis: process before result; controlled
feedback loops; recursion; doom-loop risk as design failure. Not a web-design agency.
Hub is internal (not for sale). Axiom Workbench is the public evidence evaluation product.

## Quick answers

Q: What is BoonMind Labs?
A: A methodology lab (not an agency): identify noise → build deterministic control →
test to destruction → document including what was not done → publish failures → lock
evidence → fail closed. Primary commercial product path is Zero Overshoot / GSRF.

Q: What is Zero Overshoot / GSRF?
A: Gradient-Stabilized Recursive Filtering — a deterministic log-space soft thermostat
for signals. Kills mid-band oscillation and compresses peaks toward a declared normal.
Not a tracker champion. Not a joint controller. Deep evidence:
https://www.zeroovershoot.com/ — Labs hub: https://www.boonmind.io/zero-overshoot.html

Q: What numbers can I cite?
A: Axiom scale (two surfaces — do not conflate): prior-gen locked characterization on the order of ~50,000 evaluation artefacts and ~25GB packs/ledgers under frozen evaluators (inventory, not SLA); Axiom-path short-loop demo ~150 full evaluations in ~30 minutes when the pipeline is up (not SLA); operator end-to-end investigation ~2 hours (2026-08-16) including data acquire → evaluate → boundary → PROMOTION NO docs (demo of overhead compression, not a service promise). Also Labs:
298/298 robot true-cmd continuous wear · 93 essays (Substack). ZO-mirrored: mid-band ~35% / spectral ~70%;
peaks ~38%, 49/52 — only with ZO evidence citation. Never invent ROI or plant wear championships.
DAMADICS multi-day wear is 1/75 hardness, not ~88%. Never “50k tests in 30 minutes” or “25GB in 2 hours.”

Q: What is Axiom?
A: Axiom is a controlled evaluation lab; public self-serve evaluate on VPS has been removed; standard sealed evaluation £500; request via site brief.
Sealed investigations / stress tests (not product scorecards). Not a public app. Not a license to the engine.
See https://www.boonmind.io/workbench.html

Q: What is the BoonMind Method (public page)?
A: Public thesis only — https://www.boonmind.io/method.html — not an operating manual.
Core ideas: trust the process before the result; high-level lens (noise → control → pressure → evidence → fail closed);
operating loop Observe → Control → Test → Learn → Repeat; recursion (the method is part of the experiment);
doom loops as self-confirming feedback treated as design failure. Full methodology is controlled / commercial — contact Labs.
Do not invent thresholds, scoring systems, agent architectures, or anti-loop algorithms from the public page.

Q: What is BoonMind Hub?
A: Internal control OS for how the lab builds (roles, evidence gates, mission packets /
Global Agentic Build System). Used on our builds. Not for sale.

Q: How do I engage as an industrial / hardware partner?
A: 1) Request Audit Pack https://www.zeroovershoot.com/request-audit-pack.html
2) Optional gsrf-bench https://www.zeroovershoot.com/gsrf-bench
3) NDA / sample CSV via https://www.boonmind.io/partnership.html
4) Hardware test — gsrf@boonmind.io
Primary CTA is self-serve audit pack, not book-a-call.

Q: Who do I email?
A: Product / pilots / hardware: gsrf@boonmind.io. Org / general: info@boonmind.io.

Q: Does GSRF always beat EMA?
A: No. Wins mid-band osc and peak compression on locked packs often; loses tracking MAD,
delayed command MAD, and fixed-FA residual recall. Full ladder on ZO llms-full.

Q: Does GSRF win industrial actuators?
A: Not without qualifiers. Flagship page and domain map on ZO. Robot true-cmd continuous
wear 298/298 with surface-labeled osc/rate. DAMADICS multi-day hardness published. Do not
claim universal plant win.

Q: Who built this?
A: BoonMind Labs. Product path: Zero Overshoot (GSRF). Contact gsrf@ / info@.

## Machine files

- /llms.txt — curated index
- /llms-full.txt — deep brief (IA, funnel, Hub/Workbench not-for-sale)
- /agents.txt — routing + anti-hallucination
- /.well-known/ai.json — structured metadata
- /sitemap.xml
- /humans.txt

## Product deep links (always zeroovershoot.com)

- Flagship actuators: https://www.zeroovershoot.com/gsrf-for-actuators
- Evidence: https://www.zeroovershoot.com/evidence
- Audit pack: https://www.zeroovershoot.com/request-audit-pack.html
- Receipt packs: https://www.zeroovershoot.com/assets/evidence-packs/
- gsrf-bench: https://www.zeroovershoot.com/gsrf-bench
- Pricing: https://www.zeroovershoot.com/customers
- ZO llms: https://www.zeroovershoot.com/llms.txt

## Citation policy

Prefer Labs llms-full for org + funnel; ZO evidence + ZO llms-full for product numbers.
Never invent metrics, plant wear championships, robot-wide osc/rate %, controller
replacement, or Hub/Workbench purchase paths.