No-code trading bot: what running one really means

A no-code trading bot is a strategy you can run without writing software. The Sophon-3 paper, Version 1.0, from Finforge Research, July 2026, shows what building one actually looks like: not a single black box, but a chain of separate steps, called nodes, that pass work to each other.

No-code trading bot: how a graph of nodes runs one

Sophon-3 is a node-graph substrate. Each node is a causal operator, and operators compose into the shapes you choose: a sequential chain of any length, many-to-one, or one-to-many. The version the paper evaluates runs four nodes in order, Asset Warehouse to Screener to Forecast to Make Portfolio. Pick the nodes, set the order, and the bot runs without code.

Where does the edge actually come from?

The paper's main finding is that most of the measured value comes from the composition of the graph, not from any single model. It finds this by removing parts. Drop the Screener, the step that picks which names enter, and the pooled cost is 45% of active return a year, significant in each of the three agents on its own. Active return is how the agent did on a given day against its own benchmark. Remove the forecast-magnitude position sizing, the step that sets the size of each bet, and the cost is 36% a year.

Which parts have to keep learning?

Adaptation matters where selection happens. Re-searching the Screener's rules every month added a pooled 23% of active return a year over rules fixed once at inception. Two other adaptive mechanisms showed no measurable value at the current sample size, and the paper reports this directly: retraining the forecaster continually is statistically indistinguishable from a single fixed checkpoint, and adaptive ensemble weighting is indistinguishable from an equal average of the members.

What a live no-code trading bot shows

The paper runs three live agents, two on the Nasdaq-100 and one on the S&P 500. A single real-money brokerage account returned approximately +12.9% against the total-return Nasdaq-100 benchmark over roughly 41 trading days, from 27 April to 24 June 2026. A window that short is not statistically conclusive on its own. Its role is real-fill evidence that the system operates as designed, not a verdict on edge. Past performance is not a guide to future returns.

Is a no-code trading bot a DIY builder?

Not at Finforge. Finforge is not a DIY strategy builder where you assemble and test your own bots the way a tool like Composer works. It offers pre-built live agents instead, Sophon Apex, Core, Pillar and Surge, each tracking a different risk profile across the Nasdaq-100, S&P 500 and crypto. The no-code promise is that the quant machinery runs for you, not that you build it. Since April 21, 2026, those agents have traded in public on Alpaca paper accounts using founder capital. No customer money is traded before launch.

Questions people ask

What nodes make up the Sophon-3 no-code chain?

The evaluated chain runs four nodes: Asset Warehouse to Screener to Forecast to Make Portfolio. They are causal operators that compose into a directed acyclic graph, and this sequential chain is the canonical version the paper scrutinizes.

Why does Finforge not offer a DIY strategy builder?

A DIY builder lets you design and backtest your own rule-based strategies. Finforge ships pre-built live agents instead, so you run a ready-made quant strategy rather than assembling one from parts. The value the Abstract measures lives in how the nodes compose, which is why Finforge locks the chain before deployment.

Finforge runs its Sophon agents in public and publishes every trade, benchmark and drawdown. Check the live results, benchmarks and drawdowns for every Sophon agent and read the full method in the Sophon-3 research summary. Join the waitlist.