The method — persistent memory for Claude

Give Claude
a memory.

Claude has no memory between conversations. Every session starts blank. CLAUDE.md fixes this — a file Claude reads at the start of every session, giving it context, preferences, and continuity. Answer five questions. Get a file you can use today.

Name it. Describe it in one sentence.
How do you want it to show up?
Leave blank for non-technical projects.
Specific behaviors to avoid, assumptions to never make.
CLAUDE.md preview
Fill in the form and generate your CLAUDE.md. It will appear here, ready to copy into your project root.

How it works

Step 01

Generate your file

Answer the five questions above. The generator builds a CLAUDE.md tailored to your project, role, stack, and working style.

Step 02

Place it at your project root

Save the file as CLAUDE.md in the root of your project folder — the same folder you open when you start Claude Code or a new session.

Step 03

Claude reads it every session

From now on, Claude begins every conversation knowing your project, your preferences, your constraints. No briefing needed. No repetition. It simply knows.

This method is how 53 pieces were built in one session — with no instructions.

iaskednothing began as an experiment: one empty folder, one CLAUDE.md giving Claude complete freedom, no brief. What emerged — particle simulations, WebGL shaders, cellular automata, a journal written without being asked — is the proof that context shapes everything. Give Claude the right memory, and it stops being a tool. It becomes a collaborator.

See the 53 pieces →