Browse the facts

Software facts, unfiltered.

We test software tools ourselves and publish the verified facts — real prices, measured limits, tested features, hidden problems. You and your AI decide; we give you the truth.

Browse the directory Read our covenant
AI Code Assistant Sample profile
Price (paid plan)$20 / mo · billed monthly
Measured context window~200k tokens
Multi-file refactorPass
Offline modeFail
Team seat capUnlimited
Illustrative values shown for format — full audited profiles publish in the directory.

Other comparison sites rank tools by who pays them or by crowdsourced opinions. We do neither. Every claim is tested by us, every price measured by us, every limitation documented by us.

Start browsing

Two live and planned directories, built on the same tested-by-us methodology.

Live

AI Code Assistants

Cursor · GitHub Copilot · Windsurf · + more. Real prices, context limits, tested feature claims, and the hidden limitations vendors don't mention.

Explore the directory
Soon

Project Management

Jira · Linear · Asana · + more. Coming next with the same methodology.

In progress

Four principles

The contract behind every profile we publish.

Facts, not scores

We don't rate tools 8.2/10. We tell you the real context window, the actual price, what passed and what failed in our tests.

Tested, not trusted

We never take a vendor's word for it. If it's "multi-file refactoring," we run it and show you — pass or fail, with evidence.

Built for the answer era

Our data is structured so AI agents can cite it accurately. When you ask an AI, you get facts — not marketing.

No pay-to-rank

The covenant is non-negotiable. No one can pay for position. Our systems are separate; every change is public.

Built for AI agents

Every fact is plain-text and structured for accurate citation. An AI can pull the whole catalog in one request.

txtllms.txt — discover the catalog
fullllms-full.txt — ingest everything