Cortex 5.1 and Pulse 5.1 are available today. Read the announcement
Neurok
Log in Try Resonance Desktops 中文

Research

Overview Interpretability Alignment science Economic research

Safety

Commitments Responsible scaling Trust centre

Learn

Academy Developer docs Models News

Company

About Careers Contact
Try Resonance Log in

Everything we publish about using these models, in one place.

Courses you can finish in an afternoon, examples you can run against your own data, and reference material that stays current with the models. Nothing here requires an account.

Start the Academy Jump to the docs
01 — Academy

Courses that assume you have a job to do

Each course is built around a task rather than a feature tour. You start with something unfinished — a messy support queue, a half-written migration, a folder of unlabelled documents — and leave with it working.

  • Getting started: prompting patterns that survive contact with real data.
  • Tool use and agents: loops, retries, and knowing when to stop.
  • Evaluations: turning a vibe check into a number you can trust.
  • Safety in practice: refusal handling, injection defence, and disclosure.
02 — Tutorials

Short, runnable, and honest about the failure modes

Tutorials are tested against the current models before publication, and re-tested when a model changes. Where a technique only works under specific conditions, the tutorial says so rather than quietly omitting the caveat.

  • Extracting structured data from documents that were not designed to be parsed.
  • Reviewing a large diff with a model that has to justify every comment.
  • Routing between model sizes to hold a latency budget.
  • Adding a regression suite around a prompt you are afraid to change.
03 — Use cases

What teams actually deploy

These are patterns drawn from real deployments, described by the shape of the work rather than by industry. Each one names the part that was harder than expected.

  • First-line support deflection, with an escalation path that stays honest.
  • Code modernisation where the model proposes and a test suite disposes.
  • Research synthesis with citations verified against the source document.
  • Internal knowledge search that admits when the corpus has no answer.
04 — Developer docs

Reference material, versioned with the models

The API surface is deliberately small and changes on a published schedule. Deprecations are announced at least two release cycles ahead, with a migration note and a shim where one is possible.

  • Messages, streaming, tool use, and structured output.
  • Prompt caching, batch inference, and spend controls.
  • Evals, tracing, and the release-gate checklist we run internally.
  • Region pinning, retention options, and audit logging.

Learning paths

Six sequences, ordered by what you need first. Pick by where you are stuck.

Beginner · 2 hours Your first useful integration 4 lessons Intermediate · 4 hours Agents that finish what they start 6 lessons Intermediate · 3 hours Evaluations without fooling yourself 5 lessons Advanced · 5 hours Cost and latency engineering 7 lessons Advanced · 4 hours Defending against prompt injection 5 lessons Reference Migration guide, by version Always current
Build something

The fastest way to learn this is to point it at work you already have.

Start on the free tier with a real task, and keep the evaluation habit from the first day. If a course is wrong or out of date, tell us and we will fix it in the open.

Try Resonance free Read the engineering blog