Playground – Robin Rahman

RAG on your text

Paste or upload any text (a doc, a README, your resume), ask a question, and watch the pipeline: chunk, embed, retrieve, then answer, grounded only in what you gave it. The same RAG pipeline behind the site assistant.

Your text

Question

Vision: describe an image

Drop an image here, paste one (Ctrl/Cmd+V), or click to choose. A vision model describes it, names what's in it, and reads any text. Runs on Gemini's free tier; your image isn't stored.

Semantic product search

Search by meaning, not keywords. Try a vague need and watch it find the right products, the kind of semantic search I build into commerce sites.

Note: a fixed sample of 12 products, not a real store, so the cards don't link anywhere. In a real build this would be your catalog.

Try:

See this in a real store: R² Commerce

Sneaker Lab

Drop tooling for the sneaker and proxy crowd: semantic search by vibe now, with a live release monitor and proxy-health dashboard next.

Open Sneaker Lab

Semantic similarity

Type two phrases. They get turned into embeddings (vectors of meaning) and compared with cosine similarity, the same trick behind the blog search and the assistant's RAG.

Try:

On-chain explorer

Paste an Ethereum address or ENS name and read it live from mainnet: balance, whether it's a wallet or a contract, and token details. Read-only, over a public RPC.

Try:

Natural language to SQL

Ask in plain English. Gemini writes the SQL, it runs on a demo orders table (24 rows), and you get the rows back. Read-only.

Try:

Agent: watch it work

Give it a task. The agent plans, calls tools over Robin's real blog posts and projects, and you watch each step before it answers.

Try:

Want this built for you?

I build assistants like this, grounded in your own content, agentic, and embeddable. Want one for your site, docs, or store? Leave your email and I'll reach out.

Only used to reply about your project. No spam, no selling your address.

Live, interactive demos of the things I build.