Created by Percy Rojas Masgo — Condesi Perú / Qomni AI Lab Open standard for deterministic engineering calculations. MIT License · Live Demo · Paper (PDF)
Write an engineering calculation once. Get exact answers in <300ms, with the standard and formula cited. No LLM. No approximation. No setup.
Online (browser, no install): 👉 qomni.clanmarketer.com/crysl/
plan_pump_sizing(500, 80, 0.75)
→ Required HP: 13.26 HP [NFPA 20:2022 §4.26]
→ Shutoff HP: 15.91 HP
→ Flow: 31.55 L/s
→ Head: 56.25 m
| Question | LLM (GPT-4 / commercial API) | CRYS-L |
|---|---|---|
| "500 gpm pump at 80 psi, 75% eff — how many HP?" | ~15 HP (±15% error) | 13.26 HP (exact) |
| Standard cited? | No | NFPA 20:2022 §4.26 |
| Latency | 2,800–8,500 ms | 53–286 ms |
| Reproducible? | No (stochastic) | Yes (deterministic) |
| Works offline? | No | Yes (WASM) |
| Cost | $0.01–0.015/call | Free |
15–850× faster than LLM API. Exact answers. Traceable to published standard.
CRYS-L describes what to compute, not how. The CRYS-L JIT compiles your plan to WebAssembly (WASM) and executes it at near-native speed:
User query → match plan → extract params → execute WASM → return result
(2ms) (1ms) (1ms) (<1ms) (<1ms)
Total: ~5ms WASM execution
+ HTTP + JSON overhead = 50–300ms end-to-end
Compare to LLM: tokenize → 750M+ params → decode → 4,000–45,000ms → approximate answer.
plan_pump_sizing(
Q_gpm: f64, // required flow (GPM)
P_psi: f64, // required pressure (PSI)
eff: f64 = 0.70 // pump efficiency (0-1)
) {
meta {
standard: "NFPA 20:2022",
source: "Section 4.26 + Chapter 6",
domain: "nfpa_electrico",
}
let Q_lps = Q_gpm * 0.06309;
let H_m = P_psi * 0.70307;
let HP_req = (Q_lps * H_m) / (eff * 76.04);
let HP_max = HP_req * 1.20; // NFPA 20: shutoff ≤ 140% rated
formula "Pump HP": "HP = (Q[L/s] × H[m]) / (η × 76.04)";
assert Q_gpm > 0.0 msg "flow must be positive";
assert eff <= 1.0 msg "efficiency must be ≤ 1.0";
output HP_req label "Required HP" unit "HP";
output HP_max label "Shutoff HP (NFPA)" unit "HP";
output Q_lps label "Flow" unit "L/s";
output H_m label "Total Dynamic Head" unit "m";
}
CRYS-L is not limited to any domain. The six domains below are the initial stdlib — engineering starting points. Any deterministic calculation in any field can be expressed as a CRYS-L plan: structural, geotechnical, chemical, financial, legal, medical dosage, physics, thermodynamics, acoustics, fluid dynamics, environmental, astronomical — unlimited. If a formula can be written, it can be a CRYS-L plan.
| Plan | Formula | Standard |
|---|---|---|
plan_hazen_williams(Q, D, C, L) |
V = 0.8492·C·R^0.63·S^0.54 | IS.010 Peru / AWWA M22 |
plan_darcy_weisbach(Q, D, f, L) |
h_f = f·L/D·V²/2g | Darcy-Weisbach |
plan_pipe_sizing(Q, v_max) |
D = sqrt(4Q/(π·v)) | IS.010 Peru |
| Plan | Formula | Standard |
|---|---|---|
plan_pump_sizing(Q_gpm, P_psi, eff) |
HP = Q·H/(η·76.04) | NFPA 20:2022 §4.26 |
plan_sprinkler_density(area, density) |
Q = area × density | NFPA 13:2022 |
plan_hose_stream(class) |
Q_hose by occupancy class | NFPA 13 Table 11.2 |
| Plan | Formula | Standard |
|---|---|---|
plan_voltage_drop(P_kw, V, L, fp, rho) |
ΔV = 2ρLI/S | IEC 60364-5-52 |
plan_transformer_sizing(P_kw, fp, fu) |
S = P·fu/fp | IEC 60076-1 |
plan_cable_sizing(I, method) |
S from NEC tables | NEC Article 310 |
| Plan | Formula | Standard |
|---|---|---|
plan_beam_deflection(w, L, E, I) |
δ = 5·w·L⁴/(384·E·I) | ACI 318-19 §24.2 |
plan_column_capacity(b, h, fc, fy, rho) |
Pn = 0.85·f'c·Ac + fy·Ast | ACI 318-19 §22.4 |
plan_footing_bearing(P_kN, qa, ratio) |
A = P·1.10/qa | NTE E.050 Peru |
| Plan | Formula | Standard |
|---|---|---|
plan_shaft_power(T, rpm) |
P = T·ω | ISO 14691 |
plan_gear_ratio(N1, N2, T_in) |
T_out = T_in·(N2/N1) | AGMA 2001 |
| Plan | Formula | Standard |
|---|---|---|
plan_hvac_load(area, delta_T, U) |
Q = U·A·ΔT | ASHRAE 90.1 |
plan_heat_transfer(k, A, L, delta_T) |
Q = k·A·ΔT/L | ISO 6946 |
| Plan | Formula | Standard |
|---|---|---|
plan_water_demand(units, type) |
Q_daily from fixture units | IS.010 Peru |
plan_tank_sizing(Q_daily, hours) |
V = Q·t | IS.010 Peru §3 |
Beyond these 6 domains: Chemistry (reaction yields, stoichiometry), Geotechnical (bearing capacity, settlement), Acoustics (SPL, noise reduction), Financial (amortization, NPV, IRR), Medical (dosage calculation, BMI, GFR), Astronomy (orbital mechanics), Environmental (emissions, dilution) — any domain with a published formula can contribute a plan to the CRYS-L stdlib. The language has no domain limit.
Plan Runs: 1 2 3 4 5 | Avg
─────────────────────────────────────────────────────────────────
plan_pump_sizing (500gpm,80psi) 302 277 305 282 265 | 286ms
plan_hazen_williams (150mm,300m) 56 51 61 50 49 | 53ms
plan_voltage_drop (75kw,380v,120m)130 118 113 127 112 | 120ms
─────────────────────────────────────────────────────────────────
Overall average | 153ms
vs LLM:
- GPT-4 Turbo: 2,800–8,500ms (18–56× slower)
- Commercial LLM API: 2,200–7,000ms (14–46× slower)
- Local LLM 1.5B: 8,000–45,000ms (52–294× slower)
program ::= plan_decl+
plan_decl ::= 'plan_' ident '(' params? ')' '{' body '}'
params ::= param (',' param)*
param ::= ident ':' type ('=' literal)?
type ::= 'f64' | 'f32' | 'i64' | 'bool' | 'str'
body ::= (const | let | formula | assert | output | meta)+
const ::= 'const' ident '=' expr ';'
let ::= 'let' ident '=' expr ';'
formula ::= 'formula' string ':' string ';'
assert ::= 'assert' expr 'msg' string ';'
output ::= 'output' ident 'label' string ('unit' string)? ';'
meta ::= 'meta' '{' (ident ':' string ',')+ '}'
expr ::= term (('+' | '-' | '*' | '/' | '^') term)*
term ::= number | ident | ident '(' args ')' | '(' expr ')'Built-ins: sqrt, pow, abs, min, max, log, log10, round, clamp
crysl-lang/
├── SPEC.md # Full language specification (v2)
├── ORIGINALITY.md # Language originality statement
├── paper/
│ └── main.tex # Full LaTeX paper (arXiv-ready)
├── stdlib/
│ ├── hidraulica.crysl # Hazen-Williams, Darcy-Weisbach, pipe sizing
│ ├── nfpa_electrico.crysl # Pump sizing, sprinkler, cable, transformer
│ ├── civil.crysl # Beam deflection, column load, slab
│ ├── mecanica.crysl # Shaft power, torque, gear ratio
│ ├── termica.crysl # HVAC load, insulation, heat transfer
│ └── sanitaria.crysl # Water demand, tank sizing, sewage
├── runtime/
│ ├── interpreter.md # How the WASM runtime works
│ └── integration_guide.md # How to embed CRYS-L in your app
├── examples/
│ ├── hello_pump.crysl # First example: pump sizing
│ ├── hello_hazen.crysl # Pipe flow calculation
│ └── hello_cable.crysl # Electrical cable sizing
├── benchmarks/
│ └── results_2026-04-13.json # Measured latency data
└── README.md
- Fork this repo
- Write your plan in
stdlib/{domain}.crysl - Add test vectors in
benchmarks/tests/{plan_name}.json:
{
"plan": "plan_hazen_williams",
"cases": [
{
"inputs": {"Q": 2.5, "D": 150.0, "C": 120.0, "L": 200.0},
"expected": {"V": 0.141, "h_f": 0.089},
"tolerance_pct": 0.5,
"source": "IS.010 Peru, Example 3.2"
}
]
}- Submit PR — all valid plans following the grammar are welcome
Checklist:
-
meta {}with standard name + section reference -
assertfor each input -
formulafor each key equation -
outputwith label + unit - 3+ test cases from published tables
CRYS-L was designed and created by Percy Rojas Masgo (Condesi Perú / Qomni AI Lab) in 2025–2026.
The language, grammar, compiler architecture, and standard library are original works. No content has been adapted or copied from third-party tools, languages, or libraries.
Engineering formulas in the stdlib are mathematical laws in the public domain. Standards (NFPA, IEC, ACI, IS.010) are cited by name and section number only — consistent with academic reference practice. No text has been copied verbatim from any copyrighted standards document.
Copyright (c) 2026 Percy Rojas Masgo — Condesi Perú / Qomni AI Lab
CRYS-L language spec, grammar, standard library, examples: MIT License
The Qomni Engine runtime integration is proprietary. The language itself — grammar, stdlib plans, this repo — is fully open.
See LICENSE for full terms and ORIGINALITY.md for the complete originality statement.
CRYS-L is released as a fully open specification and implementation.
License: MIT — free use, modification, and integration into commercial systems.
Objective: Become the standard execution layer for deterministic AI computations.
What CRYS-L enables:
- Deterministic computation compiled to WASM (μs–ms scale)
- Physics-as-Oracle (PaO): equations as the source of truth
- Execution in browser, server, and embedded environments
- Exact, standard-referenced answers — no probabilistic approximation
Important distinction:
- CRYS-L (language, compiler, runtime, stdlib) — open MIT
- Qomni Engine (planner, learning loop, retrieval engine) — proprietary
Opening CRYS-L creates adoption. Keeping Qomni's cognitive engine closed preserves the architectural advantage. This is the same model used by Linux + cloud vendors, TensorFlow + Google, and LLVM + Apple.
Qomni is an independent research effort advancing a new paradigm in AI systems: execution-first cognitive architectures that minimize unnecessary neural inference.
Why support Qomni?
- Development of CRYS-L as an open execution language for deterministic AI
- Research on low-latency inference architectures (μs–ms scale)
- Creation of domain-specific crystals: engineering, legal, scientific
- Advancement of AI sovereignty — edge-first, backend-optional systems
What support enables:
- Faster iteration of the Qomni Planner and learning loop
- Expansion of the CRYS-L standard library to new domains
- Open tooling for developers and researchers worldwide
- Infrastructure for global deployment
Philosophy:
AI should think only when necessary. Everything else should be executed.
If you believe in a future where AI is faster, more efficient, less dependent on massive models, and accessible everywhere — contribute to CRYS-L or reach out: percy@condesi.pe
@article{rojasmasgo2026crysl,
title = {CRYS-L: A Domain-Specific Language for Deterministic Engineering
Calculations at Zero Inference Cost},
author = {Rojas Masgo, Percy and {Qomni AI}},
year = {2026},
month = {April},
note = {Open Standard, MIT License},
url = {https://github.com/condesi/crysl-lang}
}Built by Percy Rojas Masgo · CEO Condesi Perú · Qomni AI Lab Live at qomni.clanmarketer.com/crysl/
