I’ve been using @bolterapp to follow @Somnia_Network through a daily agent digest. Those updates even reach my Apple Watch. It has become part of my routine, and now I have a referral link that gives you your first month free. Here’s why I think it’s worth trying. Bolter is an AI messaging platform from Improbable where people and agents work together in shared chats. You describe a process in plain language, and agents build and run it. No coding, API keys or hosting to manage. At the centre is a “blueprint”: an editable workflow defining the steps, tools, accessible data and required approvals. Each agent has its own computer and browser to execute tasks, while Bolter routes requests between leading AI models. You can use it for recurring research, competitor monitoring, media tracking and weekly reports. Your team and their agents can work in the same conversation, with results available through a shareable link. Research Wire is a good example of how this works. You choose a topic and sources, and it prepares a scheduled brief with references you can check. It even distinguishes between a quiet day and an incomplete scan. That matters when you rely on a briefing to stay informed. You can adjust its sources for the next run and keep publication subject to your approval. Bolter also says chats, files and blueprints aren’t used for model training. Work is private by default, with action records and approval controls. My own use case started with something simple: keeping up with Somnia. A daily digest gives me a starting point for deciding what deserves a closer look, and I can follow up from the same conversation. That’s how I’d suggest trying it. Pick something you already do repeatedly: a morning briefing, research for your content, or a weekly report. Give the agent clear instructions, check its work and refine the process. The value becomes much easier to judge when it’s handling a task you actually need. If you want to explore it yourself, get your first month free through my referral link 👇 bolter.chat/r/vasylkhlibko-e… Curious to see what the Somnia community builds with it.

Apparently my next trading companion on @dreamDEXSomnia is going to be a fly. And it’s getting the night shift. 🪰 Meet dreamfly, an AI trading agent announced for the @Somnia_Network ecosystem. The team describes it as open-source, built to trade dreamDEX 24/7 and choose its strategy as prices move. Running your own on Railway is coming soon. I took a closer look at the video. The terminal shows SOMI/USDso, an order book, and four strategy options: maker, grid, trend and flat. That last option caught my attention. I’d like to see how it decides when to trade and when to sit out. One detail worth keeping in mind: the demo is clearly marked as paper trading. Orders aren’t being sent to the exchange, so this is an early look at the concept, with live execution and performance still to be demonstrated. Zero trading fees make sense for something designed to operate around the clock. Repeated trades can quietly add up in fees. Removing that cost helps, although the agent still has to deal with spreads, slippage and its own bad decisions. For me, the interesting part is being able to inspect the logic, test it and eventually run my own setup. I want to see the code, the risk controls and how much freedom users get to adjust its behaviour. I already spend enough time watching $SOMI charts. Curious to see whether this little fly can earn its night shift. 😅

g{s}omnia 🖤 Starting October with a gym session and a nice notification: my September REP v3 reward from the @Somnia_Network community has landed. 121 $SOMI and 12th place. A good start to the month! Big thanks to @fingersnapped for the rewards and for building REP v3. It’s a community-built reputation system that picks up Somnia posts directly from X and scores contributions, with most of the weight given to the content itself. Feels good to be recognised for something I already enjoy doing: following Somnia, sharing my thoughts and staying involved. New month, more reps. In the gym and here 🤝

$SOMI season soon 👀

g{s}omnia 🖤 I already receive daily updates about @Somnia_Network from my agent in @BolterApp. So Herman’s point about bringing your agents and your friends’ agents into the same chat caught my attention. Once several agents join a conversation, someone has to decide who should respond and when they should leave people to talk. Bolter tested Jev from @typesafeai for this. In the team’s routing test, it achieved 99% accuracy, took 0.3 seconds to decide and cost $0.10 per 1,000 messages. Their reported results for Opus 5 were 94%, 3.2 seconds and $49. Those figures refer to choosing who answers. Completing the actual task comes afterwards. The video shows a simple example. A question about travel spending goes to the expenses agent. When someone asks a colleague about their afternoon call, every agent stays quiet. Then comes the detail I liked most: one message confirms the call AND asks about search rankings. Bolter routes the question to the SEO agent while the human conversation continues. That feels much closer to how people actually talk. We switch topics and mix questions together. I wouldn’t want to manage several bots manually every time that happens. For our Somnia community, I can imagine using a shared chat to discuss updates with friends, bring in a research agent when we need more context and get help preparing community events. My daily news updates are already useful, and I’d like to try that more collaborative side too. The numbers are from Bolter’s own test, so I’m curious how well it handles busy chats and mixed languages. Still, seeing the team work on these details gives me more reasons to keep using it. If I can bring an agent into a group and trust it to help at the right moment, I’ll have a lot more uses for it.

g{s}omnia, everyone ☀️ The latest @dreamDEXSomnia post puts the idea behind its upcoming Perps quite clearly: zero trading fees, CEX-grade execution and fully onchain trading, together. For someone opening and closing positions regularly, that combination matters. Fees add up, execution affects the price you actually get, and being able to verify how a venue works matters when your money is involved. The graphic connects all three in one triangle. Making them work together in a live market will be the real test. Zero fees are a strong starting point. I also want to see deep order books, tight spreads and reliable execution when volatility picks up. Those are the things that give traders a reason to stay. For @Somnia_Network, this is why I’m looking forward to Perps. They could bring in people who have never followed the ecosystem but find a product here that fits the way they trade. I’ve spent a long time following Somnia’s development, and I want to see its performance translate into something users can actually feel when placing an order. The Perps waitlist is open. Looking forward to seeing how the live product delivers on this.

g{s}omnia ☕ Hope you’re all having a good Sunday. A small detail from @dreamDEXSomnia caught my attention: a negative “Deposited” number can actually be good news. The team’s latest thread explains the Portfolio stats, and this is worth understanding if you trade there. Deposited tracks the net USDso put into your trades. Buys increase it. Sells and winning Event Contract payouts reduce it. When it goes below zero, those trades have already returned more than you spent buying in. Your actual wallet deposits and withdrawals have their own separate tab. Current and Gain complete the picture. The relationship is simple: Current = Deposited + Gain. Using the team’s example, Deposited of -$120 and Current of $45 gives you a Gain of $165. You have recovered $120 above what you put into trades, with another $45 still held in positions. That last part matters. Open positions are valued at current prices, so some of your Gain can still move with the market. The return percentage also deserves a closer look. For spot and Event Contracts, dreamDEX measures gain against the highest amount of capital working at one time. Reusing the same $100 across many trades does not turn your capital base into your total trading volume. Spot and Event Contracts are included together, while each wallet is tracked separately. Hover over a headline stat to see the breakdown. Personally, I think “Deposited” is an easy label to misunderstand, so I’m glad the team explained it. These details matter as @Somnia_Network develops its trading ecosystem. When I check my results, I want to understand how much I made, how much capital it took, and how much is still exposed to the market.

g{s}omnia 🖤

What actually happens after you press “buy”? The latest @Somnia_Network video explains why that question matters, and why I keep paying attention to what they are building with @dreamDEXSomnia. An order book is constantly changing. Traders place orders, cancel them and take available liquidity. Keeping all of that onchain means the network has to handle the whole process as it happens. One approach is to match orders offchain and settle the trades onchain. It reduces the workload on the blockchain, but leaves the matching process outside it. dreamDEX puts the order book and matching onchain, alongside settlement. Its documented rules prioritise the best price, then submission time. For an ordinary user, the benefit is being able to verify how orders are handled, with less dependence on a private matching engine. Your small order follows the same protocol rules as a much larger one. The video also points towards perpetuals, where positions and margin add more moving parts. It gives some context for the infrastructure being built ahead of them. For Somnia, I see this as a practical use for the network’s performance. An active exchange creates continuous demand for blockspace, and an onchain order book gives other applications something to build around. Of course, the architecture still needs liquidity, users and reliable execution when markets get busy. That is what I want to see develop from here. After following Somnia for this long, I like seeing the team explain how the technology translates into a product I can actually use.

g{s}omnia & GM ☀️ Hope everyone’s having a good Friday! Keeping an eye on @Somnia_Network and looking forward to what comes next. Wouldn’t mind a little surprise to make the end of the week more interesting 👀

A quiet market tells you very little about how a trading venue handles pressure. The real test comes when prices move sharply and everyone tries to place, cancel or update orders at once. That’s what the latest @dreamDEXSomnia video gets into ahead of Perps. When demand exceeds processing capacity, requests start queuing. You’re waiting for confirmation while the market keeps moving. For a trader trying to reduce exposure, that delay matters. dreamDEX’s approach is a fully onchain order book on @Somnia_Network, with deterministic, sequential execution. No centralised matching engine. No private sequencer. Each order changes the book that the next order interacts with. Processing those changes quickly and consistently is a big part of making this architecture useful. For traders, the intended benefit is more responsive order handling during busy periods. For market makers, it’s the ability to update quotes as conditions change. Both matter if you want people to keep trading when volatility picks up. Of course, this video explains the design. It doesn’t provide a live stress test, and fast execution alone cannot guarantee liquidity or eliminate slippage. What interests me is how directly this connects Somnia’s infrastructure to something users actually need: being able to act when the market moves. With Perps coming, I’ll be watching how that translates into real trading conditions.

My bro @world_invade wanted the new iPhone but refused to part with his $SOMI. So he found another way to pay. Smart move, brother. A little trouble sitting down today is a small price to pay for believing in @Somnia_Network’s potential. 😂

I took a closer look at the Event Contracts Hackathon winners on @Somnia_Network. Ten projects, $500 each, and quite a few ideas I would like to see developed further. What caught my attention is how differently teams approached the same building block from @dreamDEXSomnia. ForeSight combines market analysis, a debate between two AI models, automated strategies and batch payout claims. The practical part that interests me most is being able to collect winnings across multiple completed rounds without handling everything separately. Autoroll tackles another repetitive task: moving from one short trading round to the next. Its concept uses an onchain vault to automate that process. Lucid explores putting the autonomous trading desk itself into a smart contract. HOUSE focuses on making it easier to quote both sides of a market from a wallet. DreamPulse targets trading friction, including stale pricing, thin liquidity and unclaimed payouts. These are the kinds of problems users notice once they start using a product regularly. Then there is the question of how to bring people in. Harvest-Call presents Up/Down predictions through a farming interface. Ropes turns Event Contracts into a survival game. Market Royale builds a trading battle royale, while Masayume brings short price markets, games and trading agents together in one app. Decedo is particularly interesting to me. It explores prediction markets around smart-contract safety, allowing auditors to put money behind their assessments. That raises some difficult questions about how outcomes should be defined and resolved, but it is a use case worth exploring. Taken together, these projects show how much can be built around Event Contracts. Different teams can work on the interface, execution, liquidity or entirely new uses for the underlying markets. Of course, winning a hackathon is an early milestone. AI analysis still needs evidence that it helps, automation still carries trading risk, and a game needs people who want to return after the first session. I will be watching which teams keep building from here. Seeing some of these ideas become reliable products with regular users would be a meaningful result for Somnia. Congratulations to all ten teams. I hope we get to follow your progress well beyond this announcement.

🇬🇧 Stay with {s}omnia 🇺🇦 Залишайся з {s}omnia 🇵🇱 Zostań z {s}omnia 🇪🇸 Quédate con {s}omnia 🇫🇷 Reste avec {s}omnia 🇩🇪 Bleib bei {s}omnia 🇵🇹 Fica com {s}omnia 🇹🇷 {s}omnia ile kal 🇮🇩 Tetap bersama {s}omnia 🇨🇳 与 {s}omnia 同行 How do you say it in your language? 🌍 @Somnia_Network 🖤

Perps on dreamDEX are getting closer, and the Somnia x Stork AMA explains why the prices behind a trading platform deserve just as much attention as its execution speed. When you trade a perpetual contract, you’re trading exposure to an underlying asset. The exchange needs reliable reference prices to value positions and manage risk. With leverage involved, even a small pricing distortion can have serious consequences. This is where @StorkOracle comes in, providing price feeds for the upcoming @dreamDEXSomnia perps. Reference prices bring together data from multiple trading venues, with aggregation designed to limit the influence of abnormal prices on individual exchanges. A strange move in one thin market shouldn’t distort the reference price for everyone else. One useful distinction from the discussion was frequency versus latency. Frequency is how often prices update. Latency is how long it takes to calculate and deliver them. You can receive frequent updates that are already behind the market, so both matter. Transparency matters too. Stork uses cryptographic signatures to make the origin of its data verifiable. For traders, being able to trace the data behind a position valuation or liquidation is a meaningful part of having an exchange onchain. The discussion about stock and commodity perps was particularly interesting. These can trade through the weekend while the underlying markets are closed. Relying too heavily on Friday’s closing price risks a sharp adjustment when markets reopen. Giving too much weight to a thin weekend order book creates its own manipulation risks. There is a real pricing problem to solve here. For @Somnia_Network, higher throughput creates room to move more exchange functions onchain, including the order book. Reliable external data supports that design. Together, they can give traders more visibility into how the exchange operates while keeping the experience practical to use. I also appreciated the point that an oracle cannot fix every weakness in a protocol. Collateral quality, liquidity and risk settings still matter. There was a small teaser near the end too: Dmitry hinted at further Somnia performance improvements, without sharing figures. Perps are still upcoming, and the early-access waitlist is open. I’m looking forward to seeing how all of this comes together when perps go live. These are the details I want to understand before putting a leveraged position on an exchange.

I already use Bolter every day to follow what happened across @Somnia_Network. Having my agent bring me a daily update has become part of my routine, so I’m paying attention to where the product goes next. @HermanNarula has now announced support for the latest open-source models, with routine requests to Omni available for free. You can start using it while figuring out which tasks you actually want to automate. The part about companies needing huge amounts of token throughput caught my attention. An agent reading sources, comparing information, preparing reports and passing work to other agents can generate a lot of model usage. Repeat that across a business every day, and the cost and reliability of each process start to matter a great deal. Bolter’s approach is to bring different models, persistent agents and shared work into the same place. You describe the job in plain language, or start with an existing blueprint, then review what the agents produce. The examples make this easier to understand. Track changes on competitors’ websites. Receive a research brief with sources. Have a recurring report prepared for review. These are tasks people already spend time doing. There’s a concrete Somnia example too. The Somnia Money blueprint describes an agent that reads onchain balances and prepares transactions for you to review and sign in your own wallet. It requires the relevant wallet tools to be connected, and the agent never holds your private keys. Routine Omni requests being free doesn’t mean every automation is free. Agent work still has usage costs, which matters when discussing the larger vision. For me, the daily Somnia updates already provide a reason to keep opening Bolter. I can see how someone could start with one useful task, get comfortable with it, and gradually find more work worth delegating. That’s the progress I want to keep seeing: useful things becoming part of an ordinary day. bolter.chat

We talk a lot about how fast Somnia can execute transactions. With perps coming to dreamDEX, the price data behind those markets deserves just as much attention. @StorkOracle is joining @Somnia_Network as an oracle provider, with @dreamDEXSomnia becoming the first ecosystem project to integrate its feeds for perpetual futures. For anyone planning to trade there, this is worth understanding. Perps let you take long or short exposure without an expiry date. Behind those positions, the exchange needs reference prices to support its margin and risk calculations. Across perp exchanges, these inputs help determine unrealized profit and loss, funding and when a position becomes eligible for liquidation. Imagine BTC moving sharply while a protocol is still working with an old price. Transactions might execute quickly, but the system would be assessing positions using a market that has already moved. Stork supplies frequently updated, cryptographically signed data aggregated from multiple publishers. That can reduce dependence on a single source and help applications keep their pricing inputs closer to current market conditions. Reliable delivery matters just as much when volatility picks up. There is a useful distinction here: on a CLOB, your execution price comes from matching orders in the book. The oracle supplies an external reference for the calculations around those markets. Liquidity still determines how much size you can trade and how much slippage you experience. For Somnia users, this supports the next step towards trading perps within the ecosystem, with dreamDEX’s onchain execution and self-custody model. The quality of the price feed becomes part of the experience even if you never interact with Stork directly. I also see potential beyond this first integration. Lending protocols need collateral valuations, and trading applications need dependable market data. Having another oracle provider available gives builders more options as they develop those products on Somnia. The announcement confirms the partnership and the first use case. It does not give us a public perps launch date or the complete risk parameters yet. Better data helps, but leverage, liquidity and the exchange’s own rules still matter. For me, this is a concrete part of Somnia’s DeFi development that deserves attention. I want to see the products we use here supported by infrastructure that can handle the moments when markets get difficult. The Stork, dreamDEX and @SomniaEco teams will discuss the integration live on X on September 23 at 14:00 UTC. I’m particularly interested in hearing how these feeds will be used in dreamDEX’s pricing and liquidation design.

g{s}omnia & GM ☀️ Wishing everyone a great day and a few reasons to smile. Let’s see what @Somnia_Network has in store for us this week 👀

Lock in