Subscribe
Share
Share
Embed
Interviewing indie founders about their journey and their products. itslaunchday.com
Dagobert Renouf (00:01)
Hey, Pramod, welcome to lunch day.
Pramod (00:04)
Here I go.
Dagobert Renouf (00:11)
Yeah, we were talking on Twitter and you really, like I think you're just getting started, right? Like you don't have many customers yet. Like how is it going so far?
Pramod (00:26)
⁓ It's hard. ⁓ I've been doing this for like some time now. I mean, I'm I know like the dynamics of ⁓ building, you know, products independently ⁓ here on the X community. But ⁓ yeah, it's hard, still hard for me because I don't I don't know like if the ⁓
the profile matters or not but I see myself not there yet where I just build something and you know do very good with the distributions I'm still figuring out so it's definitely hard. Yeah I like I just I
Dagobert Renouf (01:08)
And when did you start?
When did you start the...
When did you start building your own things?
Pramod (01:22)
I think I got into this community in 2020, early 2023s. That's when I figured out there are a lot of people like me who are passionate about building something and also can sell it and make a living out of it. So that was like my point where I was like, okay, maybe I can also do this. that's that point I realized.
that I can do this. 2023 I like explored like the space like how people are doing it or what's the process, how they build, how do they sell, how do they approach the customers, everything. Then in 2023 I was actually working in 925 back then. in 2023 December I thought
Dagobert Renouf (02:21)
Mm-hmm.
Pramod (02:21)
⁓ I would just go all ⁓ in building ⁓ public. So I quit my job in 2023 December and then I started building. Yeah.
Dagobert Renouf (02:34)
wow.
So did you have some savings? Do you do freelancing on the side? How are you making it?
Pramod (02:44)
Yeah, so I had like a ⁓ five months runway. That's the max I had. ⁓ I was like a little not much, but little. ⁓ I was not sure if I should do that, but anyways, I wanted to take a call. I did that. ⁓ But a good thing happened is that I built something called MotionShot in 2024. ⁓
March I guess. So that actually made some money ⁓ for me and then it added up to the runway and it was ⁓ it was okay like till 2024 ⁓ July August it was fine and then things went really really different. So yeah.
Dagobert Renouf (03:37)
So where was this first product? So basically your first product had some success, that's cool. Can you tell a bit more about that?
Pramod (03:46)
Yeah, so technically it is not first. ⁓ I had tried multiple things in fact, ⁓ but ⁓ I mean that comes to my mind because I like that was the first one I did something good and people purchase like I made some money out of it. So, ⁓ so in fact, it was like ⁓ third or fourth, third, maybe third, sorry.
Dagobert Renouf (04:05)
And what was this product?
And what was this product?
Pramod (04:12)
So this is like a guide making tool like you any screen based workflow you can turn it into a guide like you install the Chrome extension ⁓ you do the process and it will capture the things and it will make a guide out of it and you can host it and share the links. So yeah that was the product.
Dagobert Renouf (04:32)
Like what kind of guide? Will it be a video? Is it going to be a step by step PDF? Like what is going to be the guide?
Pramod (04:40)
So the primary focus was around video. ⁓ You can just point out like this in this particular screen. First step is this particular thing and it will put the mouse cursor. It will zoom. It will add voiceover and then you can export as PDF as well. So ⁓ in summary that was the product.
Dagobert Renouf (05:03)
That's awesome. Is it still live? Do you still sell it?
Pramod (05:06)
Yeah, yeah, I do. I mean, we'll talk more about it, but yeah, it is it is still live. In fact, I revamped ⁓ it recently as well. So, yeah, that was my.
Dagobert Renouf (05:19)
So I know
it's not your main product, but can you show us quickly? I'm just curious because that sounds interesting also.
Pramod (05:26)
Yeah, I will definitely
show. Let me share my screen.
Dagobert Renouf (05:29)
Yeah,
sounds good.
Pramod (05:42)
So yeah, this is the product. mean, as I said, I revamped it, I rebranded it to something else. ⁓ But otherwise, the guide roll, yeah. So, but when I started it like...
Dagobert Renouf (05:49)
guide role. ⁓
and can you remove
the small window? Do you mind closing the window where we see myself? Yeah, thank you.
Pramod (05:58)
Yeah, yeah, yeah, sorry, Yeah,
Right. So this was the product like you make guides using the Chrome extension. You can use it for, you know, hosting like a help desk kind of thing, helpful for onboarding processes like for HRs in any company and, you know, creating SOPs. Every company would have
Dagobert Renouf (06:24)
What's SOP? I don't
even know what SOP is. ⁓
Pramod (06:28)
⁓ It's a standard of procedure. you have, when you have a process, a procedure to be followed, generally any company would have its standard of procedure. you can turn generally what people do is like they, ⁓ they create a PDF kind of thing with, you know, using like so many tools, which, which would not be so clear as well. ⁓ So using this tool, can, ⁓
Dagobert Renouf (06:37)
Yes. ⁓
Pramod (06:58)
exactly capture the screen and say hey first step is to click this second step is to fill this detail third step is to you know hit this button and so on. So you can make this
Dagobert Renouf (07:10)
So that's kind of
like Loom, but like dedicated for processes. So you have all these extra features of making a PDF or a full guide and not just like a video. Yeah.
Pramod (07:15)
right
Correct,
that is correct, yeah.
Dagobert Renouf (07:23)
That's awesome. So like how did it... So I guess now you switch focus to another product, but it seemed like a really cool thing. So what happened with this? Like I guess you made some money and now it kind of like stopped growing? Like what was the thing?
Pramod (07:28)
Yeah, yeah.
Yeah, yeah,
so I will just stop my screen for now so that I can come back.
Dagobert Renouf (07:44)
Yep.
Pramod (07:45)
Yeah, so yeah, I started it ⁓ in 2024, February, March, I guess. ⁓ Then I sold a lot of LTDs lifetime deals. ⁓
Dagobert Renouf (07:59)
Yeah,
good idea at the beginning to get some cash. ⁓
Pramod (08:01)
Yeah, so
I somehow like there was some tipping point where like my posts went viral, Reddit, somewhere else. And then I somehow like unknowingly, you know, with the community, someone posted it on ⁓ Facebook groups. There are so many groups for LTDs. I didn't know this. I didn't know this back then.
Dagobert Renouf (08:25)
yeah, I've
learned this from someone who said you have like crazy groups, like people are just like looking for LTDs and they're like gonna buy and buy.
Pramod (08:29)
Yeah.
Absolutely, LTD is
like a crazy, that's what I figured out. Like these are all, they were new to me. I mean, even now, like it's not like super experience, but they were very new. So someone posted it on some LTD specific group on Facebook. And I reached out to them, they reached out to me. I said, can you just post it? They posted it and it went super viral. It went super viral. ⁓ It made some good money, like.
Dagobert Renouf (09:00)
Amazing.
Pramod (09:01)
By then I had some sales, ⁓ same LTDs, but that was like a tipping point and it was just, ⁓ for next two, three months it just did really good from the channel. Yeah. So I mean, there is downside. Yeah.
Dagobert Renouf (09:16)
That's awesome. Wow, cool. Okay, so you had that, you had some sales. Yeah.
And did people use it? Because I've heard a lot of people buy LTDs and they never use it. They just like are addicted to having all these deals. Yeah.
Pramod (09:27)
That is true.
That's actually, mean, it might be offensive from the other side saying addicted, but it is actually they're very kind of addicted. So they just purchase like they, whenever they see a tool with LTD, they just go purchase it. No matter if they have an immediate requirement, they just think they will use it sometime in future. ⁓ But in my case, ⁓ so I have a discard channel where, ⁓
Dagobert Renouf (09:49)
Yeah.
Pramod (09:58)
All paid customers, I, you know, in all the emails that I send, I tell them, please join this discord channel for like updates. Like I just post whatever I build. So I think there are around 80 to 90 people right now on the discord server. So they actually use it like regularly. There are few people who use it very regularly. Like for example, some agencies, right agencies, they always deal with clients and they always have something to be explained.
Dagobert Renouf (10:14)
Mm-hmm.
Pramod (10:28)
are delegated.
Dagobert Renouf (10:30)
Yes, makes sense. Yeah, you know, to me, to be honest, it seems like the kind of product that can make money in a growing market. So I'm curious, why are you focusing on something else now?
Pramod (10:31)
So they keep using it.
So, ⁓ so back then, I don't know, like it was completely my mistake. It is not like a calculated move or something. I was very new to these things. I didn't know like the importance of ⁓ giving it some time, focusing on one thing. Maybe I over, ⁓ you know, thought about myself. I don't know, like something happened then at that point I thought maybe I ⁓
point where the sales were down and the feedback ⁓ it was coming but somehow I just I don't know what happened but I just moved on to something and from then I got distracted I definitely got distracted I wish ⁓
Dagobert Renouf (11:28)
But you know,
it's also kind of understandable because you started and quickly you got lots of sales from the lifetime deal. And then, you know, which is kind of like a boost. then you go back to normal, then it must be a bit like, ⁓ okay, you know.
Pramod (11:35)
Yeah, yeah, that is another... yeah, yeah.
So
I think that is the mistake with, I mean, I understand like for even now I would say like for someone who is starting off, I would definitely say go do LTDs, few LTDs to start with because you understand the process of, you know, getting payments, speaking with customers and stuff like that because LTDs are easier to sell, at least in my opinion. ⁓ But there is downside to it is that there is no recurring, ⁓ it's not just recurring.
Dagobert Renouf (12:04)
Yep.
Pramod (12:12)
revenue but it is not there's no recurring you know pressure on you that you have to yeah you have to keep speaking with people and you have to make sure that there is no churn in LTT there is no churn technically right but there are other ways of churning but ⁓ yeah that's the downside like you there is no regular pressure on you that okay keep speaking with customers keep pushing them make them happy keep make
Dagobert Renouf (12:17)
Motivation. Yeah.
Yeah, yeah.
Yeah, it's easy money,
which is good, but then yeah, it can be. Yeah.
Pramod (12:43)
Yeah,
so that I think, I mean, I don't blame that completely. I'm sure like there were problems with me as well. ⁓ But I think that was the problem with LTDs and somehow I carried away with something else. I wish I was stuck with it, but whatever happened is happened. I tried to... Yeah, I do. Yes. Yeah.
Dagobert Renouf (13:03)
but you still seem to be maintaining it, like you rebranded it, so you're still like, it's still working and
everything, but it's just like, you were trying something new also, yeah.
Pramod (13:13)
Yeah, so I'm just trying to, you know, find someone if someone can the new one, the new branded one, because I want to focus completely on and I'm focusing on completely on crawl chart. So I'm just trying to find someone either in my circle or someone because it's a very good ⁓ tool that I see a lot of value in it. So yeah.
Dagobert Renouf (13:34)
Exactly, that's what I said. I'm like, this seems like a cool, this is like a product that can make money. So I'm not surprised. Yes, so now you're trying to sell it. Cool. Yeah, I guess it's harder to sell because if you just have LTDs, then it's not recurring revenue. So you cannot say I have, you know.
Pramod (13:40)
Yeah. Yeah.
Yeah, I tried it.
Yeah, people, that's what people, the backed off. had put it on ⁓ platforms, but ⁓ lot of people showed interest initially, but then they figured out it is all LTDs and they back off. Of course, like they have a good reason to back off. So, ⁓
Dagobert Renouf (14:08)
Yeah, yeah,
it's just that it's less valuable. It doesn't mean it's not, but it's like less impressive than if it's a... Yeah. Wow.
Pramod (14:13)
Yeah, yeah. So that's when I understood
⁓ no matter what, ⁓ no matter how easy or hard it is, it is bad to continue keep just selling LTDs. I'm not saying should not sell, but selling only LTDs is not a good way.
Dagobert Renouf (14:33)
So for this other product, you're not doing LTDs I think.
Pramod (14:38)
Yeah, yeah, no.
Dagobert Renouf (14:41)
That's very interesting, know, that I never heard that perspective that super interesting that LTDs, yes, they bring you money, but it also has some downside, especially in terms of motivation and long term building. So now you're actually not doing it. That's really interesting. Wow. OK. But at least it allows you to learn and to get some money, you know, to found the rest of your journey.
Pramod (14:55)
Yeah. I mean, you can make tons of money. you can.
Yeah, that is there. So that's the entire point of,
yeah, that's the entire point of platforms like APSUMO and stuff. They promise you like very good ⁓ sales. And I'm sure like it does. And people I see, I spoke with a lot of people who made really good money. ⁓ A lot of like, even like top products. Recently I've just figured out a few like top products that I know they, it seems like made LTDs.
Dagobert Renouf (15:14)
Yeah.
Pramod (15:30)
and sold some kind of one time payments. So it is definitely a value addition like to as a maker for you like it's a very good ⁓ you know how do you say like it's a very good support like in initial days that you have good money so that you can focus on you know but you should you should understand that it is just that like you cannot just run entire thing on
LTDs are one time payments. It is just to give you like initial push and you should definitely stop it and you should definitely focus on the service recurring service. Yeah.
Dagobert Renouf (16:08)
So there's just like a counter argument to that is because like basically if you don't have recurring revenue, then you need recurring acquisition. And that's harder. But for example, in launch day number two, there was Simon Hoyberg who does FounderStack and it's an LTD. But he has a huge YouTube channel. had hundreds of thousands of views per day. So selling an LTD is perfect for him, you know.
Pramod (16:38)
Yeah, absolutely.
Dagobert Renouf (16:38)
because he
has recurring distribution. So there's a way...
Pramod (16:42)
No, is absolutely that is very beautifully put. Yeah.
I think that is a summary. That's the line. Like if you have very good distribution, even if your ticket size is very small, go do LTDs or maybe one time payments, you make a lot of money because you have very good distribution channel. But on the other hand, if you don't have, then you just focus on like very few clients, make sure they are paying whatever price you put and you keep them happy and you keep making money. Yeah.
Dagobert Renouf (16:59)
Yeah.
Pramod (17:12)
It's process,
Dagobert Renouf (17:13)
Yeah, so, yeah. Okay, so you've done that, and then what happened that made you create this other project that you're gonna show?
Pramod (17:24)
Yeah, so I mean it starts with my mistake that I moved on, but anyways, whatever happened has happened. Then I, so as I said, like there was at least four or five months gap. ⁓ I mean time where I was like so depressed because I tried so many things, nothing was working out. ⁓ The runway was like almost done. was, in fact, I remember like a month where I was like almost done. Now I don't have money. So.
I should definitely do something else. There is no other way. So I was so, and I have a family. have to make sure it's not like I'm independent. I do whatever I want. But yeah, so that point I thought maybe I should just keep, I mean, I understand like I want to do SAS and stuff in long run, but at least to keep me going to, you know, keep the ball rolling, I just need to do something. I, so the same place where I was to.
Dagobert Renouf (17:54)
Yep.
Pramod (18:22)
doing nine to five, I got on a contract based work. It's like very, like four days a day. Yeah.
Dagobert Renouf (18:29)
Your previous job, went... Your previous job,
you started freelancing for them? Yeah.
Pramod (18:35)
Right.
So, I mean, very like just half a day, maybe three days a week, something, something like that. ⁓ because I have very good relation, like very good people. In fact, because of them, I'm like, whatever I am, learned so much. I worked there for seven years. So huge respect to them. And they have very good respect on me as well. So yeah, I,
Dagobert Renouf (18:55)
And that's good because you
know everything so with just a few days of work a week you can bring a lot of value. Yeah.
Pramod (18:59)
Yeah.
Yeah,
yeah, yeah, right. So, yeah, so I got on that and that made me a little, you know, like not go so crazy. Then I got into MVP agency as well. Yeah, I did some MVP projects like I used to take, I would have done at least four or five. I don't exactly remember in that four to five months. I, so yeah.
Dagobert Renouf (19:14)
Yeah, I hear you.
Yeah. So you made a small
free, basically small agency building MVPs for other, for other startups like coding and designing.
Pramod (19:35)
Yeah, right.
Right, everything. mean, see, it's just two weeks MVP. So there is no point of discussing code, all these things. It's just like you have idea, I'll give you running URL, done, right? So it's very simple because it's just two weeks time and I used to charge, I think 3K or something, very, nominal. So that was, I mean, that helped me a lot.
Dagobert Renouf (19:49)
I see, see. Simple. Just to test. Yeah.
Pramod (20:02)
And then I did freelancing so I am so much into the emotion community. I don't know if you know the emotion. It's a ⁓ video.
Dagobert Renouf (20:10)
Remotion, I've
heard about it. Let me check. Wait, I'm starting to do this podcast more Joe Rogan style where I actually share my screen when I look at things. So let me just look at Remotion and I'm gonna share my screen so you're gonna show me at the same time. So Remotion, let me just share my screen.
Okay, remotion make videos programmatically. Interesting.
shit, yeah, create videos with React. I heard of that years ago, but it seems to be much better now. It seems to have evolved so much. Wow, my God, JavaScript is everywhere now. We even need JavaScript for videos. Fuck my life. my God. Yeah. But this is awesome. Wow, this is... Yeah. Yeah, but it's still cool though. You know, I'm just trolling, but it's cool. Yeah.
Pramod (20:46)
Exactly.
Yeah, absolutely.
So that was the exact reaction I was like, sh...
very cool yeah so yeah
Dagobert Renouf (21:14)
Let me just see
if they have... Okay, watch demo. Let me just... shit. No. Compose with code.
Pramod (21:22)
So all those videos that you see here
are made by Remotion itself with React.
Dagobert Renouf (21:29)
And so like it's made in React and then it export as an actual video file. ⁓ Like this is not like JS, like this is an actual video, this thing.
Pramod (21:33)
Absolutely MP4 you get MP4
Right, that's correct.
Dagobert Renouf (21:42)
actual mp4 captions wow
Here in review. my God. This seems so cool. Wow. I want to do this now. I could do so many things for launch day with this. Wow. This is beautiful.
Pramod (21:58)
Yeah, in fact, MotionShot itself
was made on Remotion. ⁓ So I was like a lot into the Remotion community.
Dagobert Renouf (22:08)
Yeah, so sorry. So you were saying you were in a emotion. OK, so what were you saying about emotion? I'm sorry. I just wanted to give some context.
Pramod (22:16)
Yeah,
no, so I was saying that I was like doing a lot of remote and related stuff, not from development side, like from remote and library development, but consuming it and building tools like out of it. And it was it's a very young like Johnny the maker, like I know him very good. So it's a very young like they're very cool young people, you know, made this really cool tech. In fact, I love it like very, very good.
Dagobert Renouf (22:34)
Yeah.
Pramod (22:46)
in my opinion. So yeah, so I started like building products around it and somehow like I kept talking with them. So I became very familiar with the Remotion ⁓ tech itself. So they listed me as a Remotion expert and I started getting freelancing from that side as well. Like people need Remotion templates, people need some project based on Remotion, people...
Dagobert Renouf (22:47)
I agree, it's awesome.
Pramod (23:12)
They want me to set up some remotion on their infra aws or whatever
Dagobert Renouf (23:16)
You know,
it's interesting because usually freelancing is annoying for me because I don't need too much. But when I'm learning a new tech, I actually like doing it. You know, like you're learning a new technology and you can use freelancing to keep learning and make money. And that's really, you know, do you relate to that? I really like that. Yeah.
Pramod (23:29)
Right
Yeah.
Yeah, I do.
Yeah. So yeah, I was saying that, I mean, that was another channel where I was getting a very good amount of freelancing work as well. did ⁓ multiple freelancing work for re-motion related stuff as well. So all this kept me going. ⁓ Yeah. So again, like January, I mean, meanwhile, I tried to reconnect with re-motion shot, my first product, first as in like the product that made some money, but
somehow I had moved on and I could not go back. So I was experimenting, experimenting again ⁓ in this year, early this year. ⁓ So back then ⁓ the MCP was very hot. Like I think they released it in mid January this year, if I'm not wrong. And chart bots were there. Yeah.
Dagobert Renouf (24:31)
Yes.
So just to give people context, just to give context,
MCP is a custom, I mean, it's just APIs basically, but that you can plug with your AI tool. So like you can plug it with Cloud or with Cursor or with Chagy. No, Chagy doesn't do it, I think. And you can plug a custom server which is gonna give you specific answers to a specialized topic. So.
that for example, I have one MCP for building Launch Day that is related to my code base, because I built Launch Day in Elixir and Phoenix, and it's a Phoenix dedicated MCP that allows cursor when I code or cloud to connect to the database, make some special queries, like it's basically giving more powers to the AI. That's basically what it is.
Pramod (25:24)
Right, yeah, so I mean, ⁓ I like the idea of MCP and people were a lot on cursor and back then windsurf, I don't know what is the state of it. But anyway, so one thing I observed is that ⁓ tools, ⁓ for example, cursor, ⁓ they know your code because it's there, it has access, it reads. ⁓
Dagobert Renouf (25:54)
Yeah.
Pramod (25:54)
But
for example, if I'm ⁓ using re-motion and building something, the code that I have, yeah, it reads, it understands, but it has very ⁓ limited knowledge about re-motion itself. The documents, the API, the technical stuff, what are the functions available, when should I use this, what does this do, what are the arguments that I need to pass, so many things. I mean, it might have some knowledge, but it would have been outdated.
Dagobert Renouf (26:12)
Yeah.
Pramod (26:24)
right because it has a cutoff.
Dagobert Renouf (26:25)
Yeah,
it's usually outdated or it's gonna make mistakes, gonna use the wrong version. Yeah, I have that all the time, yeah.
Pramod (26:30)
Right.
Right. So, I was just trying to, you know, connect these two dots like documentation. Remotion has documentation, but cursor doesn't know it, but cursor supports MCP already. So I was trying to put Remotion documentation as MCP server so that you can connect with cursor and do it. So that when you ask it, for example, agentic mode in cursor,
like now it has tools to read the documentation as well. So it goes, reads the documentation, understands these functions, understands the context better. Then, you know, it in a better way. That's the, so I was just experimenting like a, yeah. So I was experimenting like a, just like a laboratory. don't, I did not have any product kind of thing. I was, I mean, I was doing something else in AI space. I was trying to build something, but
Dagobert Renouf (27:15)
Exactly.
Pramod (27:29)
This problem, I was like stuck with it. I was like, ⁓ maybe this is very interesting. I I love technical problems. So I just kept solving it, kept solving it. And I showed it to people, few people. They're like, wow, this is good. This is amazing. So what it ultimately did is you give URL of Remotion, for example, Remotion. It goes, fetches all the things. I mean, drag was already there. ⁓
I don't know, like if I have to explain, it's a retrieval augmented generation. It's like ⁓ the AI models, they themselves don't everything, right? You have internal documentation. They don't know about your internal documentation. So how do you make your AI to read the internal documentation? Is using this RAG workflow. The workflow has a few steps. First step is to take the knowledge base, index them.
Dagobert Renouf (28:17)
Yeah, okay.
Pramod (28:27)
in a particular way, you chunk them, you index them, you keep them like in a way so that you can quickly grab what is required, what is ⁓ optimal for the particular context.
Dagobert Renouf (28:36)
Yes, okay. Formatted for the AI
to be able to use it. Yeah, okay.
Pramod (28:41)
Right so
you chunk them, you index them, store them. So this is where vector databases come into picture. These are called embeddings. You take a document, chunk them and turn them into something called embeddings. Right. And you store these embeddings into something called vector databases. I mean I can explain even technical things but I don't know where to stop so I will just.
Dagobert Renouf (28:58)
Yes.
No, it's okay.
No, no, it's okay. I think we shouldn't go deeper. It'll be too deep. But okay, I get it. Okay. So there's this whole structure. So people put the Remotion URL. Then there's this thing happening where you ⁓ find it, parse everything, go to the older web page documentation. ⁓ That's it.
Pramod (29:10)
Right, exactly. So, right.
right.
Right. So now I have all the documentation.
Now I have put MCP server ⁓ in front and now, yeah, so now like in any, like even cloud, for example, they, because they only started MCP stuff. So yeah, so ⁓ rag was there. So I was saying rag was there, but I was trying to put rag over MCP. It was not a huge step. It's just experience. It's like a better experience. I'm sure a lot of people.
⁓ did that I
Dagobert Renouf (29:54)
Yeah, it's
like a small add-on that makes it better, yeah. ⁓
Pramod (29:57)
Right, right.
So I showed it to people like people were like this is amazing.
Dagobert Renouf (30:02)
So is
that what you built in the end? Is that the product you have now?
Pramod (30:06)
Right, that is how it all started, Crawl Chat. ⁓ So the name itself, was like, initially I was like, I'm crawling and I'm chatting with it. So it's like, okay, Crawl Chat, simple. But it is a lot beyond that.
Dagobert Renouf (30:19)
And
can you show us and take us a bit through it? Because it's going to be easier, I think, to get it like this.
Pramod (30:27)
Yeah, so let me share my screen.
⁓ So, ⁓ yeah, so this is what crawl chart I'm showing ⁓ the landing page. I will show the internal pages as we go. So yeah, AI chart bot for your documentation and support. I mean, this is like the current version, but when I started, as I was saying, I was just trying to put technical documentation, ⁓ you know, in the rag, whatever, whatever, and then give them APIs.
Dagobert Renouf (31:08)
So it seems
like it evolved quite a bit. So it's not just an MCP now. It's the actual ⁓ product with a tool that people can put on their website super easily. So it pivoted, kind of.
Pramod (31:13)
right correct
Yeah. Correct.
Yeah, so I don't know like how good, bad, I have no idea, but yeah, P-words were there, are there, everywhere, wherever I do, things will, I don't know, it stabilizes maybe over the time, but yeah. So, but the crux is same, like the engine is same. ⁓ It is still a, you know, a rag system, you know, which does so many things.
Dagobert Renouf (31:46)
Yes.
Pramod (31:48)
Right, ⁓
Dagobert Renouf (31:50)
But
it shows you have good product sense, I think, because you can go from... Because what you said to me was very technical and this is very user-friendly, the landing page. So you can bridge that gap very well, I think. That's really cool.
Pramod (32:03)
Yeah,
I guess so. I'm not sure, but yeah, so I'm trying that I'm sure definitely I'm trying to be as close to the customer language than using some jargons. ⁓ Yeah, so I mean the crawl chart. Yeah, so crawl chart itself has like to keep it in simple terms. It has like ⁓ three steps, ⁓ right? So.
Dagobert Renouf (32:19)
And yeah, go ahead.
Pramod (32:31)
Just first step is to, ⁓ you know, build your knowledge base. This can be either you just give some URL of your documentation or you upload files, right? Or you connect some, let's say, GitHub issues and so many other things, maybe notion pages. So this builds the knowledge base. And now you have ability to chart with it, right? When I say you, it is not just someone who is signing up to me.
⁓ It is the company generally, you know, companies sign up to products like this. They have so many clients, customers and whatnot. So they can expose this entire documentation to them so that they consume it in a very natural way. So that's where they can integrate this as chart bot either on, you know, your ⁓ website ⁓ or you can
Dagobert Renouf (33:11)
Yeah.
Yeah.
Pramod (33:29)
install it as a discard bot or slack bot. Right. So then, then the third step is to analyze your data documentation. Right. So now people are asking questions. Now, how much of the documentation that you have is answering the questions they are asking. Are they asking the questions that you are expecting or yeah. So are
Dagobert Renouf (33:36)
Yeah.
that's interesting. And how do you figure
this out?
Pramod (33:59)
Yeah, I'll explain. Yeah, so there are so many so you can find data gaps like, ⁓ you know, how much like what is the score, for example, for each questions that they're getting, what is the average score? Right. ⁓ Now, is it going up over the time? Is it going up, going down? What what group of your knowledge base is being used more to answer your questions? Where can you fix like for some question it answered wrongly? How can you fix it?
So, I tried to put the analytics part in the center of this entire thing so that you understand your audience better, your own documentation better.
So yeah, at high level, this is what crawl chart is right now. From the point that I just told as a MCP, you know, a small tie like MCP or documentation over MCP to what I have now.
Dagobert Renouf (34:57)
So after this we started having some problem with the recordings and we couldn't figure it out so sadly this is the end of the recording with Pramod but you can check out his website www.crawlchat.app and on the launch day you will also see a full demo of his product that he's recording. So I hope you enjoyed this chat with Pramod and I'll see you on the next interview.