I have talked about a little bit Project

Raven and I've been sharing my progress

in some of the videos about architecture

and today I want to show um a little bit

more about it and also something that's

a little bit more visual uh because this

project will be something that I will

launch for everybody to be able to

actually use to build their own

strategies and everything is going to

make more sense when I show this

presentation and I will be doing a demo

and showing you how it's going to look

like.

So, Raven is I don't know if that's

going to be the final name, but uh it's

what I call it now, Project Raven. It's

not just a trading bot, but think like

N8N like framework for building,

testing, and deploying trading bots. So,

I don't know if you're familiar with

N8N, but it's this automation tool where

you can create pretty complicated

uh flows with these nodes. And this is

the idea behind Raven doing that. But

for trading, but

there's no coding required to build your

own strategy. You can do it entirely

through a graphical interface.

So, it's basically a no code trading

builder

and now it's demo time. So, let me just

jump straight to it and show how it

looks like.

So, it's still pretty basic and I'm

going to explain what's happening here.

So, there is an interface on the front

end that is this node builder. So you

can see over in left um left hand side

here we have data sources

and I can put it over here and you can

see our poly market is pre-selected. In

the future I'm going to make the option

to add more exchanges. So it doesn't

need to be prediction markets. It can

also be um in this case and maybe

opinion and other platforms that we want

to connect to.

So in this case I'm going to do poly

market because that's the only thing

that works right now. But you know I'll

add more. We can uh have we have

different options for market discovery.

So I'll do just a specific market by the

slug. And in this case I'm not using

real data. I'm just going to put in

anything.

And we can also add more markets like if

we want our trading bot to connect to

several different markets. That's also

possible.

And um it's possible to add more

exchanges as well. Let's say that we

want to do Poly Market and we want to do

Kelsey.

You know, maybe you want to do a cross

marketplace arbitrage bot. You know, you

will connect to the Kelsey 15-minute

market and you're going to connect to

the poly market 15-minute market

and then build your strategy around

that. In my demo here, I'm going to do

it very basic because it's not done yet

and it's kind of just a demo. A lot of

the stuff works, but there's also a lot

of stuff that doesn't work and I'm

building it uh every single day.

Okay. And and everything I'm talking

about here is kind of it everything ties

into all my videos in the past because

this is basically what I've been

building. I just been building the core

architecture and framework. We have the

poly mark here. So, let me demonstrate.

So, we have different signals. It can be

all sorts of signals. I'm going to

demonstrate the price signal. So maybe

our strate our strategy just looks at a

price update. So I put a price and I can

connect these two. So the price then uh

subscribes to the poly market data

source

and um we can then select the data

field. So this is basically looking into

the order book in this case here. So let

let's say we want to look at the best

ask price from the order book

and then we can select the token side

and let's say in this case we want to do

the up token and all of this is

something I'm going to change because it

has to kind of auto to discover like is

it even a token maybe on a different

platform they call it something else but

this is so this is kind of hardcoded a

little bit. Um and then we can do a

condition so we can connect different

signals. This is just one signal. We can

do many different signals if we can do

some kind of a condition. So let's say

we want to do a threshold. This is what

I'm going to demonstrate in in this

example.

Let's say we want to make a trade when

the best ask price is

a certain uh price. So we do the

threshold

and we can say greater or equal than

let's say 6. So when the token price is

6, that's when we want to do a trade.

And that's where we can use our action.

I'm just going to check if you can

actually see it. You can. So we can drag

over the buy. And the cool thing here is

it will give you notification about the

errors. And you can also see down here

the error lock if uh if there's some

error you have to fix. And then you can

connect this one that you want to omit

an action when this uh threshold here is

is true.

and it's going to autoop populate the

market up here in uh in the market. I

can also choose a different market. I

don't know why you would want to do

that, but it is pos possible. And then

it's going to ch take the same token

side that you chose before, but again,

you can also change that. And there's

other parameters. For example, you can

choose the strategy type that's more

about how much do you want to buy for.

So, for example, in my case, I just want

to buy for a dollar. But I can also do

like a percentage of the capital. Maybe

I want to do 5% of all the capital and

other you know criterias and um and um

let me see and then you have different

metrics for risk management and the

order timeout and so on and so forth. So

now I have a very basic strategy where

we connect to the poly market. We look

at updates to the best ask price. We

check if it's above or equal to 0.6 and

then we buy. Then you can actually test

the run here with mock data.

So if you select mock data, we can see

how uh the events are going to come in.

And this is basically events from the

poly market pipeline that we that

basically run through the the system. So

it's it's going to be easier to buck fix

and uh really tweak your strategies.

And then another neat thing is we can

actually see the data here. So it's kind

of a snapshot. So you can see here the

live uh data that we came in the bit ask

and mid price and then you can see all

right this is what we got in and this is

the signal that triggered it. So when

was it triggered I mean what what data

was it triggered on. So you can see at

the value 0.569 which is the ask price

and then you can see here the the the

condition.

So it did not omit this as true. That

means that the the buy did not execute.

You can see at the result it's false

because 05069 is less than 0.6.

And you can also see here in the lock

that that yeah it always evaluate the

condition but it never execute the

trade. So obviously if if it was um if

it was less than let's say we do the

threshold.3 let's try run it again and

hope it's not going to break now that I

do this live. So we do the mock data

start test.

Yeah. So you can see here that now it

evaluated to true

and the action was triggered and then it

says all right at this price we purchase

for $1.

So this is sort of the the idea behind

the uh behind Raven a note like strategy

builder where you can you know start to

build pretty complicated strategies

without using any code whatsoever.

I most likely will make the ability to

add some code. Maybe you want to create

your own signal that's very, you know,

you know, specific to your use case.

Then then you can do go and do it with

an input and output. And obviously

everything that you see here in the

notes is just for demo purposes. There's

going to be a lot more that we can

choose between. Obviously, this is

something that I will work on more. So

the kind of the idea is we have the

front end which you saw now and that

will connect to an API and then when you

have constructed your strategy actually

what I can do is I can just copy paste

this so you can see kind of how it works

behind the scene

and I can uh paste it and then I have a

a basically a it's actually just a JSON

that defines the strategy then that what

happens is

uh when you click you know go or

whatever It's going to feed that JSON

over to the bot and it's actually going

to create that's at least my idea for

now. It's going to create its own

instance.

So, uh when you create your bot, your

bot will actually be on its own like

server instance and it's not going to be

with other people's butts just for

reliability and stuff like that. Um but

I will go more into detail about

deploying because it's still kind of

fussy and it's just again an idea other

plan features.

So it will be possible to depo deploy

several bot instances and this is kind

of why I illustrated with this. So you

have the data ingest and you you can

actually go and make a lot of different

butts. So when I copied the the payload

here the JSON you can take this you can

go and create a new but copy paste it

and then you can just change some values

and then you you deploy it or you back

test it or you simulate it for a week

and then if it's cool all right you

deploy it

and then uh parameter parameterize your

strategies. So let's say that all right

I have a nice strategy but I want to

change some values I just want to change

this condition I want to run the same

strategy in three separate instances all

right I can do that and then one

instance is 0.3 and then 0.4 4 and 0.5

and then you can kind of, you know,

test very fast, very efficient. There's

going to be some portfolio and strategy

level risk management. So, like I showed

over here, if we go over to the buy,

there's some default values that's just

set on your portfolio and you can also

set it specifically for the strategy.

And but this is a part that's going to

be built out more. There's going to be

back testing and live simulation. So

I'll be I will make the the possibility

to feed it a lot of you know past data

but also actually just run it in the

background as a live simulation

and then see how well it does and then

if you have been running it for a week

or two weeks and all right I'm pretty

comfortable with this all right click

click go and then it's it runs

and then the strategy you build is your

own it's not something I will sell to

other but I will facilitate you being

able to sell your strategy to other

people this is at least an idea that I

have. So, I'm thinking like a

marketplace where you can if you have a

strategy or have a nice idea, you can

sell it to other people. It's an idea.

I'm not sure if I'm going to do it, but

you know, I'm just thinking a lot of

ideas. So, potentially a a marketplace

where you can actually buy and sell

strategies

and then obviously the possibility to

ingest data from multiple data sources

and venues,

you know, crossplatform arbitrage. So,

Kelsey, Poly Market, Opinion,

potentially also Binance, crypto

exchanges, uh, Interactive Brokers, a

lot of different sources.

The price, I don't know. It's to be to

be announced, but my idea is I will I'll

be do doing some kind of a fixed monthly

cost of this. There's not going to be

any trade fees, a profit share because I

think if I add trade fees, it's just

going to completely ruin the edge of,

you know, the the bots. doesn't make any

sense because I know I know a lot of

other bots they have some kind of a

trade fee so I won't be doing that and

then uh I don't think I will be doing

profit share uh either so I think it

will be monthly cost but I don't really

know it's still very much um early and

I'm probably going to do some kind of a

free premium version maybe I'll do like

you know all right before you deploy it

it's going to be free and after when you

deploy it maybe it's it's uh paid

something like that don't know still

kind of work in progress cany So far, in

my opinion, the coolest part was the

drag and drop UI. Yeah, 100%. I mean I

mean that is the like the the cool thing

about it. That's kind of what I feel

like is a game changer because I haven't

seen anybody doing this. There is

trading bots already that have these

predefined, you know, tra um strategies

and stuff like that that you maybe you

can change some parameters, but here you

have a drag and drop builder and also

the ability to deploy a lot of

strategies, test a lot of strategies and

then you can f because that's probably

what you need to do. It's you probably

need to deploy 20 strategies to just

find one that actually works. And I

think I think for sure like newbies but

also intermediate of course if you are

high high high level you have the

resources to doing something crazy of

course. So this is kind of I think this

is probably yeah for new beginners and

then intermediate. I I am talking with

also a professional trader that's

interesting in in in this. So I think

it's it's uh not for institutions but it

is for like solo people that want to do

this

timeline. Um yeah like I hope I will

have a demo ready uh for some few select

people to uh use it within a couple of

weeks. I'm going to probably start with

like two or three people that I am in

close contact with have them trying it

out building some strategies and then uh

simulating it. And then after that's um

getting more people in more people in

and yeah and of course if you want to be

notified about this there's also a

Google form where you can put your

email. I was working on a landing page.

I'm not ready with it but I will deploy

that soon. But in the meantime if you

want to be notified when this is uh you

know when it's open then leave your

email also down below in the Google form

for the timeline. As as I said,

hopefully testing this within the next

few weeks. Hopefully have a very

early.01

version ready in a few weeks, maybe a

month.

Weight list. Yeah, this one is not

hopefully I will add this I will have

this website on running soon. So, in the

meantime, it's the link down in the

description that you use for the weight

list and then maybe in a day or two you

use this link here.

Yeah. again to keep in the loop just go

to our discord. We have uh a lot of

people joining.

So yeah, we have uh a lot of people

every day, several people every day just

joining. So it's um pretty fun in here.

We have a lot of good stuff going on.

Even if strategy is good, sometimes the

whole structure and how appbot operates

don't allow strategy to be profitable

and vice versa. 100%. And I think you're

right about that, Harry Potter, because

um that's why that's actually the reason

why I didn't do any strategies yet

because I want to make sure that like

the core features works 100%. Because

like I might have a profitable strategy.

You might have a profitable profitable

strategy, but it's not profitable

because it's uh run through maybe

subpar, you know, code that actually

doesn't facilitate this strategy

properly. So that's that's kind of my

thinking about it. That's why I've

really been focusing on just creating

the core parts here.

Yes, true. You're going to need to

optimize latency, but it's okay that not

nice thing about platform seems to be

the easy UI. Yeah, I mean for latency

it's going to be difficult. I think

actually especially like I don't think

this is going to be for extreme high

frequency trading like um because first

of all um like we've been in the discord

we have been testing out latencies and

the latencies

from into poly market is pretty is

pretty bad and you really you really

need to be very close to the server

really need what you to know what you're

doing and have like you know crazy setup

to do high frequency trading I think the

opportunities is elsewhere because high

frequency trading is you know there's

institutions doing that you want to you

know compete against them um so I think

for this is more like not high frequency

trading it's of course it's going to be

fast but not within like 10 20

milliseconds maybe 100 milliseconds yeah

so you get 130 150 so so when I did it

it was like all over the place it was

all over the place I can show you the

Discord in the Discord I added a script

So to a githrap here to measure the

latency and and the way the way I did it

it's it's it's a vcoded script. Okay. So

the way I did it connect to the the

websocket and then compare the time

stamp from the events to the time that

we received them. One issue about that

is there might be like uh clock shift or

whatever you call it like their server

time might be different from my server

time and we don't know

actually. So the so so actually the

problem about that is when I tested it

to begin with I got like uh minus

latency. So the best I got was these

results here. So I tested 2,000 events.

The median latency was 35.47

milliseconds. The mean was 44. The

minimum was six and the maximum was 173.

So it's between 6 and 173. So it's like

all over the place. And I don't think

this is reliable to do any high

frequency trading if if it's like that.

Um and and yeah, other people have kind

of the same. You can see here this guy

here, he have uh total events 100, the

medium median was 200, minimum was 60.

And um yeah, yeah, you could let the

user set the cloud computer and server

location. Yes, I think that's also a

very good idea. Um I think eventually I

want to

uh do it so

when you deploy your bot by default it's

going to be deployed as its own instance

on my server but make the ability that

it it's actually deployed on your own

server or something like that like if

you want to fine-tune something like

that. All right. You just get like the

basic the code, put it on your own

server and then it runs something like

that. But that's that's later down in in

the in the stages. Yeah. Like I for sure

I want to do something like that. Like

if I actually start to make money on

this. So like the whole idea about

YouTube when I started this basically

just trying to document my journey

making money online. So if this going to

be making a trading bot that would be

cool and then I want to take a portion

of that and actually fund other traders

maybe make a competition use the

platform and then uh the winning

strategies gets funding of X amount

something like that that would be pretty

cool guys it has been a pleasure talking

with you I have to go now because now

I'm getting a call and I need to go. So

uh this is not first time I'm going to

do a live this was just a test stream

and uh I enjoyed talking with you. I'm

actually happy about been so many

comments. I did not really expect that.

But I thought I was just going to be

talking by myself. So that has been been

very good. So uh I'll uh be closing down

now. Have a good Sunday and keep keep

building and I'll make sure to uh give

you an update. And again join the

Discord down below in the link. We have

a lot of good chats in there. And uh see

you on the next