Building a Solana Trading Bot in Public: A Transparency Roadmap

A Solana trading bot roadmap usually means a graphic: a timeline, three logos, and a launch date nobody can verify. We take a different approach at SolNexus Trade. Instead of publishing a promise, we publish artifacts — the live product, the open-source code, and the exact line between what is validated and what is still in testing. This is our transparency roadmap: what has shipped on the loop, what has not, and how you can check every claim yourself.

The transparency problem in Solana trading tools

The crypto trading-bot space runs on unverifiable claims. Screenshots of winning trades. "AI-powered" strategies with no architecture behind the label. Roadmaps that reset every quarter. When every competitor says the same thing, the only thing a trader can rely on is an artifact they can open, test, or audit on-chain.

Our design response is simple: if a claim cannot be backed by something a stranger can verify, we do not make it. That constraint shapes the roadmap below. Everything in the "shipped" section is live on solnexus.xyz today. Everything in the "in validation" section is labeled as such, on purpose.

What has shipped: the closed loop

SolNexus Trade closes the full loop that most Solana trading bots stop short of: Detect → Score (ML) → Execute → Review → self-calibrate. Each link is live.

Detect

Two independent real-time engines run in parallel. TokenPricePoller ingests Jupiter and DexScreener data to catch pool movement across Raydium, Orca, and Meteora. WalletAlertPoller tracks on-chain whale and shark wallets via Helius RPC. Whale alerts and DEX breakout signals surface from the same raw data stream, before the social feed catches up.

Score

Detected flow runs through a four-layer ML pipeline before any alert emits. L1 is a deterministic formula. L2 is historical reinforcement that nudges confidence based on how the same signal type actually performed. L3 is the core adaptive layer: a contextual Thompson Sampling bandit maintaining Beta(α,β) posteriors per signal class. L4 is an execution-policy score that applies risk constraints. Only when the Emit Gate clears all three thresholds — confidence ≥ 50, execution_confidence ≥ 58, token_quality ≥ 62 — does a signal emit.

Execute

Execution routes through Jupiter, wallet-native and on-chain. Every live transaction is verifiable on Solscan. Paper trading needs no keys at all. For live trading, a separate dedicated trade wallet you authorize holds bot funds, with its key secured in Azure Key Vault and RBAC-scoped to SolNexus — your login wallet never holds bot funds and never shares keys.

Review

This is where transparency becomes a product feature, not a slogan. Every signal is scored on +15m, +1h, +4h, and +1d. Wins and losses both appear in the signal accountability dashboard. You can audit exactly how an alert played out and reverse-engineer what your own trade would have done from the alert tick.

Self-calibrate

When a position closes, its P&L flows back into L2 reinforcement and the L3 bandit. The posteriors re-tune automatically. Fresh bots start conservative and earn aggression as samples accumulate. No manual retraining, no repainted indicators. This feedback loop is the core moat: the system learns from every closed trade, including yours.

What is in validation — and why we say so

Live execution is feature-complete and entering final market-condition validation. It is not generally available. We do not imply otherwise.

That distinction matters. A roadmap that presents everything as "live" forces traders to guess which parts they can trust. By labeling validation status explicitly, we let you decide what is ready for your own use. Paper trading across Smart-Money, Non-Pump, and AI Curated strategies is the available product today, with parallel backtests so you can compare outcomes before committing capital.

The artifacts you can check

Transparency only counts if it is verifiable. Open these:

• Live product: solnexus.xyz — the market overview and sample alerts run in the browser.
• Architecture: solnexus.xyz/bot/guide — the loop, the pipeline, and the Emit Gate documented in plain text.
• Open source: the MIT-licensed freqtrade adapter, CI green and forkable. Read what shipped rather than trusting what is promised.
• On-chain proof: live executions link to Solscan. No screenshots.

The AI plan layer — entry triggers, sizing, take-profit, hard stops, and three invalidators — runs on frontier GPT models behind a second-pass verifier that downgrades over-optimistic output. We do not name a model version until the API serving it is confirmed, and the plan system forbids profit promises by design.

What is next, and who sets the order

The next phase focuses on broader signal coverage, deeper DEX pool analytics, and firmer wallet-tracking tooling. The priority order is not set by a marketing calendar — it is set by testers. We are onboarding traders now to test the end-to-end loop before public launch. Testers receive three months of Pro free in exchange for structured feedback at go-live. Their input shapes the post-validation roadmap.

During testing they extract real value now: live whale and shark alerts, the signal accountability dashboard spanning +15m to +1d, paper trading with adjustable confidence scores and parallel backtests, full pool and wallet analytics, and early detection of new-launch tokens most traders never find manually.

How to get in

Login is wallet-native through Phantom — self-custody, no card. The waitlist captures your email, X handle, and trader profile at solnexus.xyz/waitlist. If you want to verify the loop before we open it publicly, that is where the roadmap starts for you.