Tutorials

The module pages explain what each part of the platform is. These tutorials show you how to use it — end to end, in the order you'd actually do it, with every call given for the CLI, the TypeScript SDK and curl.

Every step is a real API call against the single public contract — there is nothing here that only the console can do.

Before you start

You need two things for any tutorial on this page:

1. A credential. A nat_sk_… API key, or a session JWT from Auth.

2. A shell that knows about it. Every example below reads NATURALI_TOKEN:

   export NATURALI_TOKEN=nat_sk_...

Pick your client once and the tabs throughout the tutorials stay on it.

Each tutorial is atomic: it delivers one complete, working value on its own, and anything it needs from earlier is a linked prerequisite rather than a repeated walkthrough.

Available tutorials

They are listed in the order a newcomer should read them — each one's Prerequisites section names what it expects you to arrive with, and links to the tutorial that produces it.

📄️Create a provider

Create a project, register a provider — naturali models (managed) or your own key (BYOK) — and prove it works with one smoke-test generation.

📄️Your first agent generation

Bind an agent to your provider, run a generation, and read exactly what the run cost.

📄️Connect a Discord channel

Connect a Discord bot to an agent and prove it answers real users in direct messages.

📄️Structured output

Constrain an agent to a JSON Schema so its generations return a validated object instead of prose.

📄️Search a library of PDFs

Turn a folder of PDFs into a knowledge collection whose passages you can retrieve by question, with the chunking that fits your documents.

📄️Ingest images and audio

Register converters so screenshots, scanned PDFs and recordings become searchable documents alongside your text files.

📄️Split a library between agents

Split a knowledge base into topic collections and wire one search tool and one agent to each, so every agent answers only from its own part of the library.

📄️A board that runs itself

Build a kanban whose column runs your agent on a card and routes it onward by itself, then read the audit trail proving automation made the move.

📄️Wait for a slow job on a board

Add a poll column to a board so it checks an asynchronous job by itself, on an interval, until it's done — no hot loop, no human relaying status checks.

📄️Erase a person's data

Satisfy an erasure request — destroy a run's recorded content, prove it is gone, and keep the billing and audit record intact.

📄️Run an agent that records nothing

Put one agent into zero-retention so its prompts, tool calls and results are never written to disk — while it keeps answering and keeps billing.