In the first two parts we collected the skeleton of the advisor: quotes, prompt, model, storage, core. Now the main question is - in what form can it be collected?
A trading advisor is not a thing in itself. And not something in a box that you can put on a shelf and forget. It needs to be launched, you need to connect to it, you need to request analysis from it, and then receive a signal.
And as soon as you say this, the question becomes practical: in what form will we use this product? Although there is a choice, it is not unlimited.
The most understandable option is to run the program on a personal computer: desktop, laptop, it’s not so important. The second option is a self-hosted solution, when we launch the service on a rented machine on the Internet. And if we are greedy or enterprising enough, choose what suits you best, you can put together a full-fledged SaaS and provide trading signals as a service.
If we simplify it to one diagram, all three options differ in their answers to the same questions: where the advisor lives, who holds the keys and data, who pays for the infrastructure and tokens, and who is responsible when something breaks. Next, let's go through each box separately and see which modules remain the same and which ones change dramatically along with the delivery form.
"Mine, everything is mine" - desktop version
Desktop is the most direct and understandable option: the advisor lives on the user’s computer. Everything he knows and does remains nearby: settings, keys, candle history, cache, signals, logs. This immediately relieves half of the worries about “where is my data” and “who else can see this.”
The advantages here are obvious. Firstly, everything is stored locally, without external infrastructure and unnecessary intermediaries. Secondly, access to signals is fast: you open the application and see what is happening, without depending on your server. Thirdly, this option is easier to expand in the future: you can add plugins for exchanges, models, news sources and not turn the product into a monolith, if the architecture allows it.
But the desktop has a price. Uptime is far from 100 percent: the laptop was closed, the computer went to sleep, the Internet disappeared, and the advisor “disappeared”. Plus the data is not small: the history of OHLCV is growing quickly, add analysis results, cache, logs, and then more news. And all this will live on the user’s disk. Finally, the load is also local: processor, memory, network, background tasks, and the more often you analyze the market, the more noticeable this is felt..
"Let it lie here for now" - self-hosted service
Self-hosted is when you rent a server on the Internet, deploy an advisor on it and use it as a service. Deployment, config, keys, updates are on your side. Access is usually via the web, sometimes via an API, and the interface opens in a browser from any device.
The advantages start simple: uptime is close to 100 percent. The server does not go to sleep and does not close the laptop lid. Further more important: data sources are closer and more stable, there are fewer random network failures, it is easier to maintain periodicity, and as a result, the signals are more even and predictable. Another practical advantage is access from anywhere: you can connect from different devices and, if necessary, give access to trusted people without sending them the entire project or the keys on a flash drive.
The disadvantages are also standard. Infrastructure costs money, even if not cosmic. Maintenance requires either skills or a person who will cover these skills: updates, backups, logs, monitoring, banal “why the process fell.” And there is the topic of security: keys and data live on the server. Usually this is not a reason to panic, but the “set it and forget it” mode works worse here than in the desktop version.
"Fresh signals! Signals to whom?" - your SaaS
SaaS is not so much a form of delivery as a form of business. You don't "hand the box" to the user and ask him to deploy the service. You maintain the infrastructure yourself and sell access via subscription: signals as a service.
However, in terms of implementation, this is noticeably different from self-hosted. There you have one server and one owner. Users, roles, limits, data isolation, and everything that usually goes along with it appear here: billing, subscription accounting, control of token costs and loads. Plus, you need not only code, but also people who will support it: monitoring, incidents, updates, customer support. And yes, this is already a conversation about processes, and not just about architecture.
But the reward is also different. SaaS is the level at which the product begins to live its own life: you can scale, quickly roll out changes, accumulate statistics on signals, improve quality and package it in understandable tariffs. In fact, this is no longer an “advisor for yourself”, but a platform with growth prospects..
Summary
The advisor does not have one correct box. There is your task, your mode of use and your limitations. For one person, a desktop is more logical, because everything is local and under control. For others, self-hosted is better, because you need stable uptime and access from different devices. And if you think about the product and subscriptions, then this is already SaaS, with its own economy and organization.
The choice should be made not according to taste, but according to plans and capabilities: how much time are you willing to spend on support, who will be responsible for uptime, where the data and keys will be located, and how you will grow if the project suddenly “flies.”
According to itprolab.dev
You May Also Like
Apple publishes new rules on cryptocurrency apps
According to TechCrunch, a number of Apple developers recently joined forces to create the “Developers Union.” The union asked Apple to provide Apple Store users with the opportunity to download free trials of their applications. This is one of the first times that independent iOS developers are trying to fight back against Apple Store policies.
Lity. New smart contract language
There are currently more than 1,700 decentralized applications (DApps) published on the Ethereum network, and their number continues to increase. And although all Dapps rely on smart contracts, the reliability of smart contracts themselves is questionable - cybercriminals have already earned more than a billion dollars from hacking them.
