Developer Update #6 - Real-Time Accuracy, Smarter UX, and a Major Leap Toward Live Trading
Jul 27, 2025
Cam - Chief Technology Officer
What’s Coming in Our Biggest Simulation Upgrade Yet
This Monday, we’re rolling out one of the biggest platform upgrades since Nexgent launched. Our simulated trading environment is evolving to become far more accurate, responsive, and informative - bringing us significantly closer to live trading support.
This update includes deep architectural improvements and several new features that directly impact how agents simulate trades, evaluate performance, and optimize strategies. Let's dig in.
🧠 Real-Time On-Chain Swap Quotes with Jupiter API
One of the most impactful changes in this update is our full integration with Jupiter's swap quote API, powered via QuickNode. This gives us access to real-time price data directly from Solana’s DEX ecosystem - including:
Actual route plans used for swaps
Slippage-adjusted expected outputs
Swap fees
Price impact
Platform fees and fee tokens
Previously, simulated trades relied on static or midpoint pricing logic, which didn't always reflect real-world execution conditions. Now, every agent decision can be made with a clear view of what the trade would actually cost or return, down to the routing path, slippage buffer, and fee breakdown.
Why Price Impact Matters
In crypto, especially on low-liquidity meme coins or volatile pairs, price impact is often one of the most overlooked factors - but it can have huge consequences on your PnL.
Let’s say your agent is trying to buy a token with limited liquidity. Without accounting for price impact, the simulation might think the agent can buy at the current market price, when in reality, executing that trade would push the price up several percent and result in significantly fewer tokens received.
By integrating Jupiter, we're now able to factor in slippage and impact percentages before the trade is simulated, providing a much more realistic result — especially useful when evaluating liquidity traps or thin books.
⚙️ Scalable Quote Sampling for Agent Execution
Here’s where it gets technical.
To maintain simulation speed and respect Jupiter's API rate limits (50 RPS max) on our current plan, we’ve introduced a sampling architecture for quote execution. Instead of hitting the API once for every agent, we now:
Sample the smallest, largest, and median swap amounts for a given signal.
Use that data to interpolate the expected outcome for all other agents on that trade.
Cache and reuse these quotes to reduce latency and preserve speed as we scale.
This hybrid model lets us preserve simulation integrity without compromising performance, ensuring even when we’re processing 100+ agents on a single signal, the platform remains lightning fast and consistent.
🧪 Simulation Accuracy = Confidence in Live Trading
This upgrade doesn’t just make the simulated environment prettier - it radically increases the confidence users can place in agent results.
When we enable live trading in the future, users won’t be guessing whether their agent will behave differently on-chain. They’ll already know how it performs with fees, slippage, and price impact baked in.
This makes the migration path to live execution incredibly smooth. The infrastructure is already doing the hard parts.
🔄 Route Details: Full Transparency for Multi-Hop Swaps
One of the most powerful additions in this upgrade is the visibility into full route details for every swap.
As you may know, swaps on Solana - especially via aggregators like Jupiter - often don’t happen in a single transaction between just two tokens. Instead, they route through multiple liquidity pools, potentially across several different tokens and AMMs to optimize for best execution.
For example, a swap from Token A → Token B might route through:
Token A → SOL → USDC → Token B
Each leg of that route can have different liquidity depths, slippage profiles, and fee structures.
With this upgrade:
You can now view the full route plan for every transaction.
We show you each AMM involved, the input/output mints, and the swap percentages across the route.
The system even breaks it down into lamport-level precision, so you can see exactly how much was moved at each step.
💡 What Are Lamports?
Lamports are the smallest unit of SOL - just like satoshis for Bitcoin or wei for Ethereum.
1 SOL = 1,000,000,000 lamports
Other SPL tokens also use varying decimal systems (commonly 6, 8, or 9 decimals), so raw blockchain data is almost always in these "base units."
In our UI:
We show you both the lamport amounts (for technical precision), and
The converted human-readable values, rounded to match each token’s decimal setting.
For instance, in the transaction detail screen you might see:
This would be rendered in the UI as:
This dual-format approach makes it easier to verify exact values, especially for power users comparing execution outcomes or reconciling on-chain performance.
And since you can now export all this data to CSV, you can build your own lamport conversion tools, compare trade routes, and even visualize which AMMs your agents tend to favor under specific market conditions.
🧰 UX Overhaul: Full Transparency Into Agent Trades
We’ve completely revamped the transaction and trade detail UI to show:
Token prices at time of trade (USD + SOL)
Real-time slippage thresholds and actual price impact
Full route details including AMMs used and mint addresses
Exact in/out amounts in both lamports and decimal format
Fees in USD and native token
Associated signal ID for traceability
This level of transparency is essential when running multiple agents or refining strategies. It allows you to see precisely why a trade succeeded or failed, and whether tweaks are required.
📤 Export to CSV: Run Your Own Analysis
For power users and analysts, we’ve introduced CSV export functionality for all transactions and trades.
With one click, you can export your full trading history - including signal IDs, swap data, fees, and value changes - and use it in Excel, Google Sheets, Python, or your favorite tool.
We want users to own their data and be able to run custom models, track portfolio evolution, or even plug it into their own dashboards.
🔥 Liquidity Traps & Burn Logic
This update also includes:
Improved liquidity trap detection using slippage, pool depth, and post-trade value changes
Enhanced token burning logic for agents that get stuck in traps (e.g. unsellable tokens post-swap)
🐞 Bug Fixes: Cleaner Balances & Fully Resettable Wallets
This release will also include fixes for two key bugs that have been impacting simulation accuracy and wallet reliability for some users.
Profit / Balance Sync Inconsistencies
Some users have noticed inconsistencies between reported profits and wallet balances. This was primarily due to a partial migration to our new signal generation engine, where trades were being handled by a hybrid of the legacy system and the new one.
With this upgrade:
All trade simulations will be fully handled by the new signal generation engine
Profit, loss, and balance calculations will be unified and consistent
Backend discrepancies between systems will no longer occur
This ensures that your agent’s profit and balance data will reflect exactly what’s happening, with no mismatches or sync issues.
Wallet Reset Bug
In some cases, resetting a wallet didn’t properly clear simulated balances or active trades. This could cause confusing UI states and inaccurate agent behavior on restart.
Going forward:
Wallet resets will fully clear balances
Any open simulated trades will be cleanly reset
The UI will immediately reflect the updated state
These fixes are essential as we continue to refine the stability of the platform — ensuring that every agent action is predictable, accurate, and transparent.
📅 Upgrade Schedule
We’ll be taking the platform offline Monday at 9:00am UTC for deployment and production testing. We expect the maintenance window to last 2–3 hours.
The platform will be under full-time monitoring and support immediately after deployment.
💡 Final Thoughts
This wasn’t just a frontend facelift. It was a deep rebuild of our signal generation engine, trade simulation infrastructure, and user experience model.
The goal? Give you the tools to understand and trust every decision your agent makes - and prepare you for a seamless move to live trading when the time comes.
Check out the updated documentation for a full breakdown - and stay sharp.
As always, we’d love your feedback.