Tools

QuantG Open Infrastructure

AI, ML & Engineering Tools

Client-side calculators, dataset diagnostic tools, and technical estimation utilities built for machine learning engineers and AI teams.

LLM & Pricing Live

LLM Token & API Cost Estimator

Calculate character/word tokenization, compare real-time pricing across major foundation models, and download CSV cost reports.

Client-Side Editable Rates
Launch Tool →
Machine Learning In Pipeline

Model Training VRAM & GPU Estimator

Calculate exact GPU VRAM needed for model training or inference based on parameter size, precision (FP32/FP16/INT8), and batch size.

PyTorch / CUDA VRAM Math
Machine Learning In Pipeline

Time-Series Forecast Metric Evaluator

Upload prediction CSVs to calculate RMSE, MAE, MAPE, and sMAPE instantly client-side without sending dataset rows to a server.

Client CSV Parse Forecast Analytics
Machine Learning In Pipeline

Classification Confusion Matrix & F1 Tool

Input True/False Positive & Negative counts to compute Precision, Recall, F1-Score, MCC, and plot confusion matrices visually.

Model Evaluation Interactive Plot
Cost & ROI Live

AI Automation ROI Calculator

Estimate annual cost savings, hours freed per employee, and projected financial return on investment for workflow automation.

Financial Planning Instant Math
Launch Tool →
LLM & Architecture In Pipeline

RAG vs. Fine-Tuning Decision Matrix

Interactive diagnostic evaluating latency, data update frequencies, and domain specificity to choose between RAG and Fine-Tuning.

System Design Scoring Logic
Machine Learning In Pipeline

Vector DB Memory & RAM Estimator

Project RAM and disk space required for storing vector embeddings across dimension sizes (384d to 3072d) and index types.

LanceDB / Qdrant Memory Math
LLM & Architecture In Pipeline

Voice Agent Latency Budget Calculator

Break down milliseconds across STT (Faster-Whisper), LLM generation, and TTS streaming to hit target response times.

WebRTC / Streaming Latency Budget
LLM & Benchmarking In Pipeline

Sub-2B Local LLM Benchmark Comparison

Side-by-side performance scorecard comparing Gemma, Qwen, and LFM models on task accuracy and local inference throughput.

Local Llama.cpp Edge AI
Machine Learning In Pipeline

Data Normalization & Z-Score Previewer

Paste raw numerical arrays to visually preview MinMax Scaling vs. Standard Z-Score Scaling distributions instantly in-browser.

Data Preprocessing Visual Distribution
LLM & Pricing In Pipeline

Context Bloat & Prompt Cache Estimator

Estimate token bloat reduction and cost savings when implementing 4-tier Markdown context caching and memory pyramids.

Token Reduction Prompt Caching
Machine Learning In Pipeline

Client-Side Browser Embedding Sandbox

Generate vector embeddings and calculate cosine similarity directly inside the browser using ONNX models via WebAssembly.

Transformers.js Zero Server Touch

Need Custom AI or Machine Learning Systems Engineered?

We build production ML pipelines, automated RAG architectures, local Llama instances, and time-series forecasting engines.

Consult an AI Architect
Frequently Asked Questions

QuantG Tools & Infrastructure FAQ

Everything you need to know about our client-side execution models, data privacy standards, and engineering utilities.

Yes, 100% free. Because our utilities run entirely in your web browser using JavaScript and WebAssembly, there are no API costs, server overheads, or artificial usage limits on our end. You can calculate costs, benchmark models, and process datasets as many times as you require.

No. Zero data leaves your device. All calculations, tokenization heuristics, CSV parsing, and model estimation scripts execute locally inside your browser session. We do not store, log, or transmit your prompts, financial inputs, or raw dataset rows.

We update our foundation model pricing matrix continuously as providers (such as OpenAI, Anthropic, Google, and DeepSeek) release new rate card adjustments or new model tiers. Additionally, our tools allow you to manually edit input/output pricing fields directly in the interface to test custom enterprise discount rates.

Yes. Beyond public utilities, QuantG engineers custom production AI systems—including specialized local Llama execution pipelines, sub-500ms real-time voice agents, automated RAG architectures, and time-series forecasting engines tailored to specific client workloads.

Our tools are built as standalone HTML/CSS/JavaScript blocks with zero external framework dependencies. You can drop the code directly into a Custom HTML block in Gutenberg or any page builder (Elementor, Divi, Blocksy) and it will adapt seamlessly to your active theme layout.