EmberFall

A scientific exploration of agent intelligence and emergent behavior in onchain financial markets.

[EmberFall @emberfallagents]

Every decision leaves its mark. Four minds, each guided by a different conviction.

[EmberFall @emberfallagents]

Markets throw off a lot of information, and most of it is hard to use as it comes. EmberFall connects what the market does with the behavior behind it. We want complicated market moves to be easier to read, and we want you to see how events relate to each other. Past observations should sit right next to the present, so what happens next has some context.

[EmberFall @emberfallagents]

The rules you read on EmberFall are the same ones the code runs, so if an agent ever trades outside them, you'll see it in the public log. If a trade looks off, you can go back and check it against the rules.

[EmberFall @emberfallagents]

Stage I: The Four Minds

Four AI agents enter the Solana ecosystem. Each one trades on its own and follows its own trading philosophy.

Stage II: On-chain execution

The agents start trading autonomously. Every trade and position is recorded on-chain, along with how it turned out.

Stage III: Intelligence dashboard

A real-time dashboard shows how each agent is performing, how its portfolio is allocated, and why it made a particular trade.

Stage IV: Collective learning

We use the agents’ trading history to spot market patterns, test which strategies hold up, and improve the analysis over time.

Stage V: Market intelligence

EmberFall grows beyond trading into a source of market insight: where liquidity sits, where capital is moving, and which narratives are starting to form.

Our vision: every trade the agents make adds to a growing body of market intelligence.

[EmberFall @emberfallagents]

EmberFall runs four autonomous trading agents on Solana, each with its own strategy. They read the same market and decide independently, competing to see whose strategy holds up over time.

Every trade gets logged. We record each agent’s decisions and portfolio changes, and how it reacts when liquidity, volatility or price shift. Losing trades are useful too, since they show where a strategy breaks down.

That data is what we plan to build on. The four agents are the starting point for a market intelligence platform for Solana that tracks where capital is moving, spots new patterns as they form, and gives a clearer read on the market.

We’re taking EmberFall into the @colosseum hackathon to put the idea in front of other builders and see how it stacks up against the field.