NativeSoul — a coding agent that’s actually yours
Founder pricing is open until launch. The price goes up after. ❯ Get NativeSoul, $29<br>❯ For developers who run more than one coding agent.<br>A coding agent that's actually yours<br>Like one developer who already knows your work, in every tool you open.<br>Get NativeSoul — $29<br>30-day money-back. Runs on your machine.<br>See how it works ↓ ~/project — one soul, every host SOUL LOADED<br># claude code<br>$ claude<br>✓ voice: terse, dry<br>✓ remembers: 14 repos<br>✓ values: correctness
# codex — same machine<br>$ codex<br>✓ voice: terse, dry<br>✓ remembers: 14 repos<br>✓ values: correctness
One agent. The same soul and memory, live in every tool you open.
Works inside the tools you already use.<br>Claude CodeClaude DesktopCodexGemini CLIGrok BuildAntigravity<br>Runs on your machine Your code & memory never leave it No usage telemetry Never copies your credentials Closed source · 30-day money-back
❯ The Problem<br>Every session, you start over.
Today<br>You paste the same context again, every session<br>The agent forgets what you decided an hour ago<br>Switch tools and you start from zero<br>It says an integration is connected, then it breaks<br>You're the one holding it all together
With NativeSoul<br>It already knows who you are and how you work<br>Your decisions stick, session after session<br>The same agent follows you into every tool<br>It tells you the truth about what actually works<br>You go back to building
You were told this is just how AI coding works. It isn’t. It’s a layer your tools never gave you, and you’ve been paying for its absence.
❯ The Solution<br>The agent you set up once, everywhere you work.<br>It’s like having one developer who already knows your work, showing up in every tool you open. Not a new hire you onboard each morning. The same one, who knows how you build, everywhere you code.<br>NativeSoul gives your coding agent a personality you shape and a memory of your work, in every tool you open. It runs on your machine, and you can remove it any time.<br>It's yours.<br>You decide how it talks, what it values, and how it works. That stays yours.
It remembers.<br>Your projects, your decisions, the conventions you set. No more re-explaining.
It's honest.<br>It only says a thing works after it checks. When something is off, it tells you.
It's safe.<br>When it runs a follow-up for you, it runs once, and you can see exactly what it did.
It rides your host’s own hooks and MCP surfaces, with nothing injected in the middle, so it never sits between you and your agent.<br>Get NativeSoul — $29 30-day money-back · runs on your machine
// same soul, two hosts<br>Claude Code<br>▸ voice: terse, dry▸ remembers: 14 repos▸ values: correctness
Gemini CLI<br>▸ voice: terse, dry▸ remembers: 14 repos▸ values: correctness
one personality · one memory · carried everywhere
❯ Setup<br>How it works.
01 Install with one prompt
Buy once, then run npm i -g nativesoul and paste one prompt into your tool. It activates with your license, sets itself up, backs up anything it touches, and checks that everything works before it says it is done. Change your mind later and nativesoul uninstall restores your host config.
install<br>$ npm i -g nativesoul<br># backed up ~/.config — nothing lost<br># verified 6 hosts — all healthy
02 Shape your agent
Answer about five quick questions: how it should talk, what it should value, how you like to work. It saves your answers to plain text files you own and can read any time (SOUL.md, IDENTITY.md).
SOUL.md<br>$ cat SOUL.md<br># voice: terse, dry, senior<br># values: correctness over speed
03 Open any tool
Open any tool you use. Your agent is already there, with the personality you shaped and the memory of your work. Same agent, same decisions, every tool.
any host<br>$ claude # SOUL loaded<br># same agent, same memory, every host
Get NativeSoul — $29 30-day money-back. Runs on your machine.
❯ Features<br>Everything you get.
One product. Everything included for $29.<br>One agent, every tool<br>▸Saved as plain files you own (SOUL.md, IDENTITY.md)<br>▸Rebuilt at the start of every session on each host<br>▸Rides each host’s official hooks and MCP surfaces, nothing injected in the middle<br>▸Every host: Claude Code, Claude Desktop, Codex, Gemini CLI, Grok Build, Antigravity
SOUL.md · rebuilt per host<br>name: "Your agent"<br>voice: terse, dry, senior<br>values: correctness over speed<br># claude · desktop · codex · gemini · grok · antigravity<br>✓ nothing injected in the middle
Memory that stays with you<br>▸Stored locally, scoped per project so nothing bleeds between repos<br>▸Captured and deduplicated automatically; promoted on review<br>▸Open the files. Read them. Delete them. They live under ~/.nativesoul/.
~/.nativesoul/memory — local files LOCAL<br>$ ls ~/.nativesoul/memory/<br>daily/ projects/ archive/<br># scoped per project · indexed in state/nativesoul.sqlite<br># deduped on capture · promoted on review<br>✓ nothing bleeds between repos
It doesn't fake it<br>▸A capability reads "ready" only after it passes a real auth and health check<br>▸Installed is not the same as ready, and it won’t pretend...