Clean up scattered rules
Classify identity, output style, safety boundaries, project behavior, tool-specific notes, and conflicts.
Independent AI coding utility
AI coding memory.
Paste your Claude Code, Codex, or Cursor rules. Coding Memory Lab normalizes messy memory, flags conflicts, generates downloadable configs, and estimates visible token savings without pretending that output compression equals total bill savings.
Coding Memory Lab is independent and is not affiliated with, endorsed by, or sponsored by Anthropic, Claude, OpenAI, Cursor, GitHub, or the Caveman project. Do not paste secrets, API keys, cookies, or full private repositories.
Memory pack generator
Build a Claude skill, Codex AGENTS.md, Cursor rule, and compact shared memory from one input.
Do not paste secrets, API keys, access tokens, cookies, production database URLs, or private customer data.
0input tokens
0pack tokens
0%memory reduction
0/100portability score
One local memory, many coding agents.
Your rules should not fall apart when you switch from Claude Code to Codex or Cursor. This workflow turns scattered memory into compact, compatible files that you can inspect before use.
Build the pack
Measure visible savings separately from bill impact.
Caveman-style output can reduce visible text. Your actual bill also depends on input context, tools, cache behavior, and subscription limits.
0original output tokens
0compressed tokens
0%visible output saving
0%estimated bill impact
Paste an AI agent response above, then run the simulator.
Built for the actual workflow pain.
The value is not another tutorial. The value is getting one clean memory pack that works across tools, stays short, and tells you where compression can hurt safety.
Export real config files
Download Claude Code SKILL.md, Codex AGENTS.md, Cursor rules, and a shared compact memory pack.
Separate claims from math
See visible output reduction and estimated total bill impact as different numbers.
What this replaces.
Manual prompts are fine once. They break when you need the same behavior across multiple coding agents.
| Approach | Good for | Breaks when |
|---|---|---|
| Plain “be concise” prompt | One-off short replies | You need persistent rules, safety escalation, and cross-tool portability. |
| Caveman-style output | Reducing visible text quickly | You mistake output savings for total bill savings or lose code-review context. |
| Manual CLAUDE.md / AGENTS.md editing | Advanced users who enjoy config work | Rules duplicate, conflict, or drift between tools. |
| Coding Memory Lab | One compact memory source exported to multiple tools | You need team permissions, history, or SSO. Those are not in the first small-tool scope. |
Free core tool. Paid audit only if you need depth.
The generator stays usable without an account. Paid intent is for larger local audits, batch exports, and clean reports.
$0
Generate memory packs, run the savings simulator, download files, and read the comparison guides.
$9 intent
Signal that you want a larger local audit, conflict report, and batch exports without uploading a repo.
No account wall
No login, no team dashboard, no private repo upload, no claim that your total bill drops by a fixed percent.
Get the local audit release note.
Leave an email if you want the next version that runs a deeper memory audit and exports a cleaner report.
Your email is used for product updates. Do not submit secrets. The local preview stores fallback submissions in your browser only.