A classified collection of AI agent tooling

Every tool, properly classified.

One specimen, each catalogued by category, pattern, and use case. Where the copper mark appears, we installed the specimen and ran it ourselves.

Most recent acquisition

No. 001

n8n

n8n describes itself as a fair-code platform for building and deploying AI agents and workflows. It can be self-hosted on your own infrastructure, including on premises or in a private cloud.

CategoryAutomation Platforms

Reading a label

The classification key

Category
Where the specimen sits in the collection — frameworks, browser agents, automation.
Pattern
How it works structurally: tool-calling loop, stateful graph, node pipeline.
Use case
The job it is actually good at, stated narrowly rather than broadly.
Copper mark
Present only when we installed the specimen and completed a real task with it.

Wings of the collection

Browse by the job, not the tool

Wing 1

Operations

01 specimen

Wing 2

Data

01 specimen

Examination method

Identical for every specimen. A tool that fails the third stage is still catalogued — it simply does not receive the copper mark.

Stage I

Source

Every claim traced to documentation, repositories, or changelogs. Marketing pages are not a source.

Stage II

Classify

Category, pattern, and use case assigned — so a specimen can be found by the job, not the name.

Stage III

Examine

Installed and used to complete one real task. Only then does the copper mark appear.

Stage IV

Re-examine

Every specimen revisited on a 90-day clock, with corrections published on the label.