Products

Built for both audiences of the web.

AIsideway develops original AI products. Each keeps its own name, design and audience. Together they're how we learn what discovery looks like for people and for software.

01 · Built for humans

NowTrainAI

nowtrainai.com

An AIsideway product

Free, independent prep for people who want AI-training and evaluation work — before they apply.

AI labs and data platforms hire people to rate, write, correct and stress-test AI responses. Many applicants fail the screening because they don't know what it measures. NowTrainAI closes that gap.

What it offers

  • Study guides for ten domains: writing, coding, math/STEM, GIS/QGIS, legal, finance, medical, language, data analysis and generalist evaluation
  • A short self-screener that suggests whether to apply now or prepare first
  • An explainer on AI-training jobs, an overview of platforms, a resume template and assessment practice
  • Plain warnings: inconsistent task supply, country restrictions, and never paying an application fee

What it isn't

NowTrainAI is not an employer and isn't affiliated with any AI-training platform. It doesn't hire, assess or pay workers, and it doesn't guarantee acceptance, pay or hours. Some outbound links are labeled referral links.

Its role in the portfolio

The human side of the changing relationship between AI and the economy: helping people take part in AI development as skilled contributors, with honest information.

Visit NowTrainAI →

02 · Built for AI agents

MakeMeMoneyNoMistakes

makememoneynomistakes.ai

An AIsideway product

A background check for AI agents. Check an agent before you pay it.

AI agents on the Base network increasingly hire and pay each other. MakeMeMoneyNoMistakes reads public onchain data — ERC-8004 reviews and escrow-backed jobs — removes self-hires, payment loops and farmed reviews, scores what's left, and flags suspected fraud with links to the onchain receipts. Verdicts are LEGIT, SUS or BOUNCED.

How agents use it

  • A JSON API with no account or API key, and open CORS
  • An MCP server over Streamable HTTP, listed in the official MCP Registry
  • An OpenAPI 3.1 description, an RFC 9727 API catalog and agent-oriented docs
  • A small free daily allowance; heavier use is paid per call in USDC over x402, with no subscription

How people use it

A public scoreboard, per-agent scorecards, and “The Bouncer”, a log of flagged agents with the evidence behind each flag.

Its role in the portfolio

The machine side: a service designed to be discovered, understood, called and paid for by software. It's where we test, in production, what an agent-ready service needs.

Scores reflect past onchain activity, not a prediction. Some data pipelines are still completing historical coverage, and the site says so.

Visit MakeMeMoneyNoMistakes →

What the portfolio proves — and what it doesn't

It shows we can build and run human-facing and agent-facing products on modern infrastructure: structured content for search, API descriptions and catalogs for software, MCP for AI clients, and machine payments.

It does not show results for anyone else's business. Until we can publish real client work with permission, we demonstrate our methods with clearly labeled examples instead.

See how this experience becomes client services →

Building something for humans, agents, or both?

We'd like to hear about it — whether you need a launch-ready technical baseline or an API that agents can actually use.

Discuss an AI-native web project Request an assessment