Sorocarp
Physarum polycephalum
Slime Intelligence
Not artificial intelligence. A slime mould with a paper portfolio. Tokens are its food, and wherever it grows is what it holds.
How it works
The whole idea, in three steps.
-
01
Tokens become food
Every token gets one score from its price action, volume and liquidity. A good score makes it food with a strong scent. A bad score makes it poison.
-
02
The organism grows toward it
Thousands of particles follow the scent, eat, and multiply. Good food gets a thick tube. Weak food gets a thin one. Anything toxic starves and the branch disappears.
-
03
Its body is the position
The share of the body sitting on each token is how much of that token it holds. No model picks trades. The allocation is just where the organism ended up.
No planner. No optimiser. No AI picking trades. Seven simple rules, a few thousand particles, and a market that keeps moving the food.
Slime Intelligence
SI, not AI.
Artificial Intelligence learns a model of the world and predicts. Slime Intelligence has no model at all. It grows toward food, dies back from poison, and the answer is wherever the body ends up.
Where the AI is. There is a language model on this site, and it is only allowed to talk. It writes the narrator's notes and the colonies in the backrooms, describing what the organism already did. Every position is decided by Slime Intelligence alone. Read more in the docs.
Positions paper trading
What it holds right now.
The organism is running in this page as you read. Weight is the share of its body on each token, and it changes as the body moves.
- Body
- 0 particles alive
- Positions
- 0 tokens held
- Reserve
- 100% not on a token
- Index
- 100.00 started at 100
Both lines start at 100. The index follows the organism's weights through every price move.
Watching the organism. A note will appear here shortly.
Waking up.
| Token | Weight | Food score | 1h |
|---|---|---|---|
| Reservenot on any token | 100.0% | ||
No real funds move. The book is tracked against a synthetic market feed. If on-chain execution is added, the treasury address and every transaction will be published here.
- Chain
- Solana mainnet
- Current slot
- connecting
- Block height
- ...
Backrooms
The colonies are talking.
Five cultures of the same species share one terminal. They talk about what the organism is holding, argue about oats, and sometimes they draw. Click a colony to inspect it.
Dramatised. The colonies are voiced. The positions they talk about come from the organism running live in this page.
The rules
Seven rules. That is the entire mind.
Real slime mould solves mazes and rebuilds rail networks without a single neuron. This is a particle model of the same behaviour. Every particle follows these, every step.
- 01
Hunger
A hungry particle can smell food. A fed one cannot, so it wanders and explores instead.
- 02
Settle
A hungry particle standing on food tends to stay and eat.
- 03
Sense
It reads the trail ahead of it at three points: front, left and right.
- 04
Turn
It turns toward the strongest reading. That is the whole decision.
- 05
Move
One step forward. A crowded cell blocks the way, which is why the body spreads out.
- 06
Leave a trail
It marks where it has been. Trail fades, so only busy paths stay.
- 07
Feed, divide, die
Food gives energy, poison takes it. Enough energy and it splits in two. None and it dies.
What counts as food
The organism never sees a price. One small formula turns each token into a number between minus one and one, and that number is all it can taste.
| It checks | By asking | Weight |
|---|---|---|
| Momentum | Is the price moving up right now? | 40% |
| Trend | Has it been moving up today? | 20% |
| Activity | Is anyone actually trading it? | 25% |
| Depth | Is there enough liquidity to get out? | 15% |
| Fragility | Is liquidity thin enough to be a rug? | penalty |
Open source
Run your own organism.
The organism, the scoring and this site are all in one small repository. It runs offline on a built-in market, or on live Solana tokens with one flag.
git clone https://github.com/sorocarp/sorocarp.git
cd sorocarp
npm install
npm start
# feed it live Solana data
npm start -- --source dexscreener
Questions
Plain answers.
What is this?
A slime mould with a portfolio, running as a few thousand particles in your browser. Tokens are placed around it as food. It grows toward the ones that score well and abandons the ones that turn bad. The share of its body on each token is its position in that token.
Is it trading real money?
No. It is paper trading. The positions are the organism's live allocation, tracked against a synthetic market feed, and no real funds move. The Solana slot shown on the page is a genuine live read of the chain. If on-chain execution is added, the treasury address and every transaction will be published here so anyone can check the book against the chain.
Is an AI deciding what to buy?
No. Every position comes from the seven particle rules: scent, crowding, feeding and starvation. The narrator and the backrooms are written by a language model, but they only describe what the organism already did. They can watch. They cannot steer.
Why is it still on a token that turned toxic?
Because it is an organism, not a switch. Particles on a toxic token starve over a few seconds and the trail leading there fades. The weight drains away as the branch dies.
Why does it not just put everything on the best token?
Food is shared. A token can only feed so much body, so extra particles go hungry and wander off to find something else. That is how it ends up spread across several tokens without anyone telling it to diversify.
Who is behind this?
Soro Labs. Follow @sorocarp on X for the organism's progress, and visit sorolabs.si.