GUIDES & PLAYBOOKS

Practical guides, templates, and technical playbooks to help you ship AI locally—from quick validations to production-ready deployments.

Deploy a Spam Detection Model Locally (Spam/Ham) — Python (CPU)

Run a spam/ham text classifier locally with Python and Hugging Face Transformers (CPU). Load a local model/ folder and validate inference on 10 French samples.

READ

1-Hour AI Mini-Test (Model vs Dataset vs Agent)

Validate an AI idea in 60 minutes using 30–100 real examples. Set safe thresholds, spot costly errors, and decide whether to start with a ready model, build a custom dataset, or move to an agent (CPU/GPU).

READ

Dataset vs Model vs Agent: Choose & Combine

A practical, non-technical guide to picking the right AI building block: an off-the-shelf single-task model, a dataset to train/adapt it to your company data, or an agent powered by a language model for complex workflows. Includes a 60-second decision path, concrete examples, and common pitfalls.

READ

Deploy French Spam Type Detection Locally

Step-by-step guide to run the model locally, expose a minimal FastAPI API, test with curl, and set safe thresholds with a 3-zone workflow (auto / review / no action) to avoid false blocks.

READ

Private RAG with Citations (Company Docs)

Answer from internal PDFs, Word files, and emails with verifiable citations—then deploy faster with the Knowledge Base QA Agent.

READ

API & Integrations

Explore API guides and integration playbooks to connect Kinoux to your tools, automate workflows, and run AI models in real use cases.

Kinoux Models API

Run AI models through a simple inference API. Browse available models, send requests, and get structured predictions in real time.

READ

Automate AI workflows with Make

Use Kinoux in Make to build structured workflows with advanced logic. Combine AI processing with multi-step scenarios and real data flows.

READ

Use Kinoux with Zapier

Connect Kinoux to Zapier to add AI steps to your automations. Test real use cases quickly with live data, without building a custom integration.

READ