Olivia
Online
Cam.exe - Nexgent.ai
@ItCompiledynza
Founder & dev of @NexgentAI Building agentic, open-source trading infrastructure on Solana.
Sydney, Australia
Joined December 2025
33 Following
124 Followers
79 Posts
https://t.co/3qLiRoqUVR
Open-Source Trading Engine v1.1.1 is live
Stability patch, infrastructure move, and streamlining development on the on the signal generation engine.
Full release notes:
https://t.co/yYYAoRRx9o
Infrastructure update: We migrated RPC providers in this cycle. The token lock mechanism was briefly offline as a result - it's fully restored and operating as expected.
On the @NexgentAI Signal Generation Engine:
I've been investing heavily into how we build - not just what we build.
Agentic workflows across the full dev lifecycle. Claude, Langflow, and custom agents handling research, spec generation, iteration loops, and testing. The feedback cycle compressed at every stage.
This space is evolving fast and we're building on the bleeding edge of it. The upfront investment will pay for itself many times over in velocity and output quality.
This is going to help us improve the Velocity strategies, and deliver on Ignition and Leviathan.
🚀 Nexgent Trading Engine v1.1.0 is live
Introducing Auto-Trade, plus major stability improvements, validation fixes, and config export/load.
Full release notes:
https://t.co/rKmul0j56F
What is Auto-Trade?
Auto-trade follows your existing DCA + take-profit logic and keeps exposure to selected tokens automatically.
You choose the tokens per agent. The engine:
- Re-enters after take-profit
- Reconciles every 10 minutes to open missing positions
- Gates buys behind optional market-cap bounds
- Respects all existing guardrails (balance, position limits, price impact)
It’s designed for larger-cap, more stable tokens where tighter DCA + TP levels make sense.
This is early / experimental - use carefully and only with amounts you’re comfortable risking.
v1.1.0 also includes:
- Fixes for stop-loss misfires after DCA
- Sim PnL stability improvements
- Position size validation (0.05 SOL minimum enforced)
- Slippage validation upgrades
- Redis + cache resilience fixes
- WebSocket stability improvements
- Structured logging with Pino
- Graceful shutdown handling
Plus…
Trading Config Export & Load
- Export your full strategy to JSON
- Load config from another agent
- Copy settings when creating new agents
No more manually rebuilding configs.
This release is a big step toward making Nexgent not just powerful - but production-grade and resilient.
More upgrades coming. 🔥
For the first 24 hours we're going to run the signal engine in a heightened mode which means more signals - these may not be as high quality, this needs to occur to ensure the engine is working as expected. The engine hasn't released any signals yet (though Solana markets are pretty dull today), so we're dialling up.
🚀 IT’S LIVE. THE OPEN-SOURCE TRADING ENGINE HAS LAUNCHED 🚀
Today is a massive milestone!
The @NexgentAI open-source trading engine is now live, and you can officially get set up and running.
This has been a long time coming - and I can’t thank the community enough for the patience, feedback, and belief that got us here.
What’s live today
- The open-source trading engine is available
- You can deploy your own instance and start configuring agents right now
- Full simulation + live trading support
- DCA, stop-loss, take-profit, moonbag - all ready
I’ve put together a step-by-step YouTube walkthrough showing exactly how to get started, from deployment to connecting signals.
https://t.co/BAgMULsEnm
📡 Signal engine status (important)
The signal generation engine is warming up.
It’s currently backfilling on-chain data following recent upgrades, so you may not see signals for the first couple of hours. That’s expected.
Once warm-up is complete, signals will begin flowing as normal.
Velocity plan access
If you already have tokens locked:
- You’re automatically subscribed to the Velocity plan
- No action required
- Just plug your engine URL into the signal UI and you’re good to go
You can set everything up now and be ready the moment signals start flowing.
More coming (very soon)
I know there’s a lot more content people want - deeper strategy guides, agent tuning, optimisation tips, advanced configs.
Those are coming over the next few days.
Right now the priority was getting everything live, stable, and usable. The education layer will follow quickly.
🚀 IT’S LIVE. THE OPEN-SOURCE TRADING ENGINE HAS LAUNCHED 🚀
Today is a massive milestone!
The @NexgentAI open-source trading engine is now live, and you can officially get set up and running.
This has been a long time coming - and I can’t thank the community enough for the patience, feedback, and belief that got us here.
What’s live today
- The open-source trading engine is available
- You can deploy your own instance and start configuring agents right now
- Full simulation + live trading support
- DCA, stop-loss, take-profit, moonbag - all ready
I’ve put together a step-by-step YouTube walkthrough showing exactly how to get started, from deployment to connecting signals.
https://t.co/BAgMULsEnm
📡 Signal engine status (important)
The signal generation engine is warming up.
It’s currently backfilling on-chain data following recent upgrades, so you may not see signals for the first couple of hours. That’s expected.
Once warm-up is complete, signals will begin flowing as normal.
Velocity plan access
If you already have tokens locked:
- You’re automatically subscribed to the Velocity plan
- No action required
- Just plug your engine URL into the signal UI and you’re good to go
You can set everything up now and be ready the moment signals start flowing.
More coming (very soon)
I know there’s a lot more content people want - deeper strategy guides, agent tuning, optimisation tips, advanced configs.
Those are coming over the next few days.
Right now the priority was getting everything live, stable, and usable. The education layer will follow quickly.
Was nice to demo the open source platform, setup and answer some of the communities questions around @NexgentAI
Replay available for anyone who missed it!
If you’re a NEXGENT holder — must watch.
If you care about AI + automated trading but want transparency and self-custody — also must watch.
This is where personal trading infrastructure is heading.
🎥 https://t.co/iSagHKO5u0
(Starts at min 00:36)
🚀 New @NexgentAI website is live!
https://t.co/VkNuAzZGnr
The product has evolved a lot since our initial launch, and the old site no longer reflected where Nexgent is today.
We’ve refreshed the site to clearly explain:
- the open-source trading engine
- the proprietary signal generation engine
- how the two work together
- and where the platform is heading
Hopefully this makes the architecture, vision, and value of Nexgent much clearer.
🚀 https://t.co/AbQXI8iV42 LIVE AMA, This Thursday!
Join us Thursday at 9:00 AM UTC for a live Ask Me Anything as we count down to the launch of our open-source platform dropping Sunday, Feb 15.
We’ll be covering:
🔹 The official launch of NEXGENT AI open-source live trading
🔹 What’s coming in the platform release
🔹 Community questions
🎙 Hosted live on X
👉 Follow @NexgentAI to tune in and join the conversation.
Spread the word, big things ahead.
🚀 First live trading results are in - and this is a big milestone for @NexgentAI
Today we ran our first live trading tests using the new Take-Profit + DCA combination.
This setup allows an agent to:
- buy into weakness
- sell into strength
- repeat the cycle automatically
- all while managing risk without emotion
And it did exactly that.
The results (kept intentionally small to control risk)
- Peak capital deployed: 1.39 SOL
- Trades individual tokens traded: 4
- Total return: +0.30 SOL
All trades were managed entirely by the agent - DCA on the way down, partial take-profits on the way up.
No emotion. No second-guessing. Just execution.
I’ll be honest - even with small size, it’s still a bit nervy watching real funds move. But I sat back and trusted the process, and it delivered exactly as designed.
A few important notes for context
- This was the first live test using DCA + Take-Profit together
- Trade sizes were deliberately small for reconciliation and safety
- Tokens had enough liquidity to support much larger trades
- Tight DCA and tight take-profit levels were used
- Profit shown in Nexgent is slightly under on-chain profit (likely due to conservative fee estimation)
- Trades occurred during an overall flat / down-trending market
What’s encouraging is that even in those conditions, the agent was able to buy dips and sell bounces profitably.
Link to transactions and summary
https://t.co/CDbURwP223
Link to wallet
https://t.co/nlsR9v9ajZ
I get feedback that I work on too many new features quite often, but there is a reason for that, I promise - proof is in the pudding. Take profit is an incredibly powerful feature, especially when paired with DCA.
AMA & Platform walkthrough
The team will be hosting an AMA and walkthrough of the platform later this week (probably Thursday) ahead of the full platform release on the 15th of February💪
Been testing the Moonbag + take-profit combo and this is where @NexgentAI will really shine.
Picked up a token around a $32.5k market cap that later ran close to $1.6M - almost a 50x move.
Now realistically, most people don’t hold for those kinds of multiples (and rightly so). More often than not, if you try, you round-trip back to zero.
This is where the new setup changes things.
With take-profit levels and more relaxed stop-losses, you can:
- Lock in gains progressively on the way up
- Reduce risk as the position grows
- Still keep a small Moonbag in case the token keeps sending.
So you’re not forced to choose between “sell too early” or “hold everything and pray”.
You bank profits… and still keep exposure to the long tail.
This feature’s only been running in simulation for a couple days while we validate calculations and execution behaviour - but it’s already doing exactly what it’s designed to do.
Really excited about what this unlocks.
Buckle up - release countdown is on!
Markets have taken a hit lately. But it’s markets like these where projects with real utility and long-term vision win.
As mentioned in my last post, we're fine tuning the signal generation engine. One of the things we've noticed is how we're getting really good entries, but leaving a lot on the table by closing out the trade too early.
Been heads down testing a new feature that’ll be live at launch - Take-Profit.
It’s basically the other side of stop-losses. Instead of only protecting the downside, this lets your agent secure profit progressively as a trade moves in your favour.
Which, if you’ve been around Solana for more than five minutes, you know is pretty important, things move fast.
The idea is simple:
- Lock in gains on the way up
- Reduce risk as the position grows
- Stay in strong trades longer instead of fully exiting too early
You can run presets, or go fully custom and define your own levels. There’s also a moonbag option, so you can keep a slice of the position in case the token keeps sending.
Running it in simulation now to make sure everything executes exactly how it should under volatility before it’s turned on for live use. Docs will get updated once it’s final.
Markets have taken a hit lately. But it’s markets like these where projects with real utility and long-term vision win.
As mentioned in my last post, we're fine tuning the signal generation engine. One of the things we've noticed is how we're getting really good entries, but leaving a lot on the table by closing out the trade too early.
Been heads down testing a new feature that’ll be live at launch - Take-Profit.
It’s basically the other side of stop-losses. Instead of only protecting the downside, this lets your agent secure profit progressively as a trade moves in your favour.
Which, if you’ve been around Solana for more than five minutes, you know is pretty important, things move fast.
The idea is simple:
- Lock in gains on the way up
- Reduce risk as the position grows
- Stay in strong trades longer instead of fully exiting too early
You can run presets, or go fully custom and define your own levels. There’s also a moonbag option, so you can keep a slice of the position in case the token keeps sending.
Running it in simulation now to make sure everything executes exactly how it should under volatility before it’s turned on for live use. Docs will get updated once it’s final.
Launch Update
We originally targeted end-of-month for the @NexgentAI open-source trading engine going live. After reviewing where we are, we’ve made the decision to set a final, fixed launch date of February 15.
Let me explain why - and why this is a positive step.
The open-source trading engine itself is complete. Infrastructure, execution logic, simulation mode, live trading mode, multi-agent system, configuration layers - all of that is built, stable, and running well.
The work happening now is focused on the signal generation engine, which is the intelligence layer that determines trade quality and profitability. This has been something that's been neglected whilst focus has been made on the trading engine side.
Over the past week we’ve made significant upgrades:
- New iterations of the underlying machine learning models
- Expanded wallet tracking infrastructure
- Deeper on-chain PnL and accumulation behaviour monitoring
- Early-detection logic for tokens entering high-growth phases
Early signs from these upgrades are very encouraging. We are identifying tokens much earlier in their lifecycle, which opens the door to significantly larger upside than before.
However, we are still seeing:
- Occasional false positives
- Stop-loss behaviour exiting trades too early on high-potential tokens. This is fully configurable in the engine, but requires fine tuning.
- A need to introduce smarter take-profit logic to capture larger moves while managing risk
In short:
The system is now better at finding big opportunities - but we want to refine how positions are managed once in those trades.
This is not about fixing something broken.
It’s about tuning the system to maximise the advantage we now have in early detection.
We are also in an extremely volatile market phase. Models must be robust across both favourable and chaotic conditions. Allowing more time now ensures the system adapts correctly, rather than launching into instability.
So what does this mean?
- The platform is ready.
- The intelligence layer is being upgraded.
- We are improving signal quality, trade management, and risk handling before going live with real capital.
February 15 is the date. No moving goalposts beyond this.
This window is about turning a good system into a high-performance one.
When we launch, we want it to represent the standard Nexgent is aiming for - not a rushed version we have to patch in public.
Appreciate the patience. The work being done right now materially increases the long-term value of the platform.
Some trades from the last 3 hours of live trading demonstrating why we need to still make some tweaks. Huge potential as you can see.
🚀 BIG MILESTONE - The documentation for the @NexgentAI open-source trading engine is LIVE
As promised, the full platform documentation is now available so everyone can explore how the open-source trading engine works before we go live.
You can now read through:
- What the OS trading engine does
- All configurable trading logic (SL, TP, DCA, stale trades, agents, etc.)
- How signals connect to execution
- Full setup guides
- How to customise and extend the platform
Where to start
Non-technical users → Cloud hosting quick-start
https://t.co/3oqZm8nnhm
Technical users → Architecture & integrations
https://t.co/gf7akMPKG7
Nexgent x Railway - Infrastructure partnership
Additionally, we've been working with @Railway to streamline your deployment process. We've created a deployment template that lets you spin up your own instance in under a minute.
This includes the full stack (frontend + backend), so there’s no need to manage separate hosting environments. Railway makes this super easy to do!
The GitHub repo is still in private mode for now, so deployment isn’t open just yet - but you can review the docs and be fully prepared ahead of release.
We’re very close now.
Final work is focused on the signal-engine UI and backend (enabling you to connect your instance to receive trading signals) and final polish.
The foundation is built.
The docs are live.
The engine is ready.
Full release window
End-of-month target, with early February as a buffer to ensure everything launches at the quality standard we expect.
It's starting to heat up. 🔥
Sneak peak into the @NexgentAI signal connection UI.
Connecting Nexgent trading signals to the Nexgent open-source trading engine takes ~20 seconds.
How it works:
- Generate an API key in your self-hosted trading engine
- Copy your engine URL + API key
- Paste them into the Nexgent Signals UI
- Done - signals start flowing immediately
No complex setup. This is exactly how it should feel.
A quick note on token tiers:
- The amounts shown for Ignition and Leviathan are placeholders for now.
- At go-live, we’ll start with a single tier, and all existing users (who have locked tokens at any rate) will have access by default.
Additional tiers will roll out later once they’re ready.
The goal here is simple:
Signals → execution → control
With as little friction as possible.
More to come.
Documentation work continued tonight, and things are coming together nicely.
Alongside that:
- Wrapped up the first round of security review
- The platform is now with a few engineering mates for a second-pass review + stress testing
- Identified and resolved an issue related to database transaction atomicity during this process
Docs-wise, the structure is now clear:
- Open-Source Trading Engine (~90% complete and almost ready to ship)
- Nexgent Signals
- Tokenomics
Once the trading engine docs are live, the remaining sections are relatively quick to finalise.
@CRYPTO_IDIOLOGY It’s a starting point, we don’t intend for it to remain that way forever as we add additional models / tiers 🙂
We’re not at the stage where we have that many traders which will impact performance. I suspect after it goes live we may though - will be monitoring closely!
One of those full weeks where life stacks up 😅
Between strategy heavy days at my full-time role (and a few late nights) as we plan for the year ahead, plus celebrating my wife’s birthday and our anniversary, progress naturally looks a little quieter on the surface. That’s just part of the journey - and an important reminder that momentum isn’t always measured day by day.
That said, still managed to get in a solid block of dev work today, with more lined up tomorrow.
If all goes to plan, platform documentation should be live in the coming week - and after that, the platform goes live.
Steady progress. Eyes forward. 🚀
Going well! We’ve reduced the number of signals the platform produces daily from ~200 to around 10-20 high quality signals which appears to be working well. This is only in our test environment at the moment as we continue to refine. Moving away from the scalping trades to trades with higher returns - found the scalping ones were a bit hit and miss with live trading.
Obviously with a change like this our models need more data to continue to improve, but it’s all happening as we speak 🙂
Most Popular Users
Elon Musk 
@elonmusk
240.2M followers
Barack Obama
@barackobama
119.3M followers
Donald J. Trump
@realdonaldtrump
111.6M followers
Cristiano Ronaldo
@cristiano
108.8M followers
Narendra Modi
@narendramodi
106.9M followers
Rihanna
@rihanna
97.2M followers
NASA
@nasa
92.1M followers
Justin Bieber
@justinbieber
90.5M followers
KATY PERRY
@katyperry
86.7M followers
Taylor Swift
@taylorswift13
80.5M followers
Lady Gaga
@ladygaga
72.1M followers
Kim Kardashian
@kimkardashian
69.3M followers
YouTube
@youtube
68.6M followers
Virat Kohli
@imvkohli
68.4M followers
Bill Gates
@billgates
63.4M followers
The Ellen Show
@theellenshow
62.5M followers
CNN
@cnn
61.9M followers
Neymar Jr
@neymarjr
61M followers
X
@x
60.9M followers
CNN Breaking News
@cnnbrk
59.9M followers