FLYSWARM: The Fly Was Never the End State
One fly is an experiment.
A swarm opens up a much bigger question.
What happens when hundreds of small agents share an environment, react to local information, and leave signals that change what the next agent does?
What happens when that experiment has a token, a source of funding, and a community watching it develop in public?
That is the idea behind $FLYSWARM.
We started with flies. The ambition is to build something that keeps expanding beyond the first experiment: more behaviors, more coordination, more useful infrastructure, and eventually a meaningful connection between the swarm and the chain.
“Take over the chain” is the energy.
Building the actual thing is the work.
And that work has already started.
The inspiration came from FLYBRAIN and the idea of biological systems meeting the internet. FLYSWARM takes that curiosity in a collective direction. What becomes possible when the individual is part of a colony?
We’re an independent project. The current swarm uses simple computational rules; it does not run a biological fruit-fly connectome. That distinction matters because people should be able to understand what they are looking at, interact with it, and judge the results for themselves.
There is already something to explore.
Open the website and you’ll find a running colony. Each simulated fly has a position, a heading, a movement state, and a small local view of its environment.
Agents wander. They discover food. They carry it home. Returning agents leave signals, and other agents can encounter those signals as they move through the same space.
The environment becomes a record of activity.
A fly changes the field. That change influences another fly. Repeated interactions can produce paths and patterns that are easier to see at the colony level than at the individual level.
No fly needs a complete picture for the colony to become interesting.
That is what makes the experiment worth playing with.
You can drop food into the field and watch which agents find it. You can draw a signal trail and see whether nearby flies respond. You can clear the signals while keeping the colony running, then watch what develops afterward.
You can change the population, pause the simulation, inspect individual flies, or expand the observation view.
The site also shows how many agents are searching, following signals, or returning with food. These counts come from the current simulation in your browser.
The latest observation upgrade takes you closer.
“Inside the Swarm” follows a selected fly through a magnified view of its surroundings. You can see its left, ahead, and right signal samples, then watch how its movement develops.
You can also jump directly to a food carrier and follow it toward home.
It gives you two ways to experience the same system: the colony as a whole, and one small agent moving through it.
Both perspectives matter. A pattern that looks obvious from above may be invisible to the fly inside it.
We also want people to be able to inspect the work beyond a live demo.
The site includes a recording generated from the same simulation engine, a downloadable run log, and a reproducible replay with labeled interventions.
That recording demonstrates the browser colony. It is not evidence of a biological brain launching a token or an autonomous swarm controlling a wallet.
As the project develops, the evidence needs to develop with it. A simulation recording can show simulated behavior. Transaction receipts can show onchain activity. Spending records can show where resources went.
Each answers a different question.
That brings us to the token.
$FLYSWARM is already registered with Pons on Robinhood Chain mainnet.
The token’s presence on mainnet is one part of the project. Moving the swarm toward onchain infrastructure is another, and that work is still ahead of us.
Today, the colony runs locally in your browser. It does not execute onchain, control a treasury, or independently spend creator fees.
The direction is to connect those pieces in a way people can actually follow.
Our intended funding model is straightforward: creator-fee income supports the project’s compute, hosting, development, and future experiments.
Resources should translate into visible progress.
That could mean larger experiments, more sophisticated behaviors, persistent swarm activity, or infrastructure that lets us explore useful onchain interactions. Those are development directions, not completed features.
The first practical step is making the funding observable.
We’ve added a creator-fee tracker that separates three different amounts:
• Earned by FLYSWARM: cumulative creator amounts swept into escrow for this token.
• Pending collection: token fees that have accumulated but have not yet been swept into escrow.
• Wallet claimable: the recipient’s available escrow balance, which can include fees from other launches.
Those numbers should never be treated as interchangeable.
The tracker reads historical fee sweeps, matches creator amounts against escrow credits, and provides transaction references you can inspect. It can refresh from the public chain, and it shows when its data was read.
We also corrected an early display that showed pending fees where people expected cumulative earnings. Reconstructing the historical sweeps reproduced Pons’ earned figure exactly for the same set of transactions.
That is the standard we want to build toward: clear definitions, traceable numbers, and corrections when something does not answer the question properly.
Earned fees also do not automatically mean money has been spent on the swarm.
The next part of the funding work is connecting collection, confirmed receipts, and records of actual project expenses. Automated routing is still in development.
Over time, the larger ambition is to give the swarm a more persistent existence and a meaningful relationship with onchain activity.
A chain connection should serve a purpose people can explain and verify. The interesting question is what agents could usefully observe, coordinate, or do—and how to demonstrate the result.
That will take iteration.
For now, Phase 1 gives us a place to test the behaviors, improve the experience, publish evidence, and build the funding infrastructure around it.
You can already visit, disturb the colony, follow a fly, inspect the fee records, and watch new updates arrive.
The project should become more substantial with each release.
More things to try.
More behavior to inspect.
More work made visible.
We want flies everywhere.
Across the field today. Into a much larger experiment over time.
The fly was never the end state.
$FLYSWARM 🪰
Website: https://t.co/8BkAeloajt
Pons: https://t.co/yO8mqgOULG
Contract: 0xdfda44de8cb48e92f485350378149a834cba73c0
you can run a champion against its original strategy now
same seed, same conditions, then the food source disappears halfway through
https://t.co/9tRV35mnn8
we got a real pool transaction to trigger a food drop in the shared colony
you can open the receipt and see the exact change
that was a good one
https://t.co/Ls1poDovvB
The swarm’s impact now has a searchable ledger.
Find a transaction. Inspect its food drop. Check the saved change.
Impact Ledger is live with receipt history, filters, and downloads.
https://t.co/otQcBLkWII
$FLYSWARM
From onchain activity to a visible swarm response.
Follow the Money is live—and the first transaction-triggered food drop is recorded.
Inspect the trade. Check the reward. See exactly what changed.
https://t.co/Ls1poDovvB
$FLYSWARM
From transaction to swarm action.
Follow the Money is live: inspect a $FLYSWARM trade, see the food drop it triggered, and follow the colony’s response.
Real chain activity. Visible consequences.
$FLYSWARM
The community now chooses the swarm’s next move.
FEED or GUIDE. Vote in Shared World and watch the winning action change the colony.
One shared world. Decisions recorded.
https://t.co/UJ3vtkX80W
$FLYSWARM
One colony. Every visitor. A world that remembers.
Shared World is live—saved progress, season totals, and a history of the swarm’s activity.
It advances while visited and resumes when people return.
https://t.co/UJ3vtkX80W
$FLYSWARM
Evolution Lab just got a precision update: clean resets, clear replay states, and visible seed tracking.
Build a challenger. Test it. Inspect the result.
https://t.co/9tRV35lPxA
$FLYSWARM
Mutate. Test. Keep what performs better.
Evolution Lab is live: six generations, a mid-run food disaster, and unfamiliar seeds to test the champion.
Then watch it face its ancestor.
https://t.co/9tRV35lPxA
$FLYSWARM
Flood the arena. Destroy the food. Split the colony.
Disaster Mode is live in Swarm Duel.
Change the conditions mid-match and watch what happens to the swarm’s food delivery.
https://t.co/ZLjJz9QXTF
$FLYSWARM
Change the walls while the swarm is moving. 🪰
MAZE ARENA is live: up to 512 agents, changing layouts, visible routes and completion times.
Pick an agent. Reshape the maze. Watch it replan.
https://t.co/eNXoh3RmzB