Starter File Generator

Set up your own memory repo. Everything runs in your browser — nothing is sent anywhere.

🚀 You’re viewing this on GitHub Pages

To use Engram, you need your own copy of this repository. Clone it to your machine, then run this wizard locally or continue here to generate starter files you can download.

git clone https://github.com/TheAlexFreeman/Engram.git

After cloning, cd Engram and open HUMANS/views/setup.html in your browser, or continue below to generate and download files.

or
📊 Browse this repo read-only on the dashboard
1About You
2Profile
3Platform

Step 1: About you

Optional — helps personalize your starter files.

Used in the greeting of generated files. Leave blank to skip.

Gives your agent a head start on understanding your needs.

🔒 This information stays in your browser. Nothing is uploaded.

Step 2: Choose a starter profile

Pick a template to pre-fill common preferences, or start blank.

💻

Software Developer

Code-first, concise, technical. Optimized for programming workflows.

📚

Researcher

Thorough, rigorous, citation-aware. Optimized for analysis and synthesis.

📋

Project Manager

Concise, actionable, stakeholder-ready. Optimized for coordination.

✍️

Writer

Narrative-first, tone-conscious, revision-oriented. Optimized for drafting and editing.

🎓

Student

Curious, explanation-seeking, building foundations. Optimized for learning.

📚

Educator

Clear, scaffolded, student-centered. Optimized for teaching and materials.

🎨

Designer

Iterative, feedback-driven, concrete over abstract. Optimized for creative work.

Skip — I’ll build from scratch during onboarding

Step 3: Choose your AI platform

This determines what instructions are generated for your first session. This wizard only creates local files; it does not edit git settings — git remote setup stays manual.

▶
Claude Code
CLI tool — already configured via CLAUDE.md
🤖
Codex Desktop
Desktop app — generate a repo-local `.codex/config.toml` for this clone
✎
Cursor
IDE — already configured via .cursorrules
💬
ChatGPT
Needs custom instructions — we’ll generate them
⚙
Other / not sure
We’ll generate a system prompt you can paste anywhere
No extra setup needed. The repo includes a CLAUDE.md file that Claude Code reads automatically. After placing your files, run claude in the repo directory.
Generate a local Codex config.

Portable config works with Codex CLI; Codex App may require absolute paths (run setup.sh in the repo).

Enter absolute repo and Python paths before generating a Codex config.

No extra setup needed. The repo includes a .cursorrules file. After placing your files, open the repo folder in Cursor and start a conversation.
We’ll generate a chatgpt-instructions.txt file. Copy its contents into ChatGPT → Settings → Personalization → Custom Instructions. You’ll share repo files manually at the start of each conversation.
We’ll generate a system-prompt.txt file. Copy its contents into your AI platform’s system prompt or session initialization.

Your files are ready

Download each file and place it in your Engram repo, overwriting any existing file at the same path. If you want a git remote, add it manually after this step.

📊 Dashboard available

Once your files are in place and you’ve run a session or two, use the Memory Dashboard to see your knowledge base, projects, and system health at a glance.