1 00:00:00,280 --> 00:00:02,740 Hello and welcome back to APIs You Won't Hate. 2 00:00:02,920 --> 00:00:03,670 I am 3 00:00:03,670 --> 00:00:06,970 mike Biko, one of the co-founders of APIs You Won't Hate here 4 00:00:06,970 --> 00:00:10,420 sitting down for a chat with my new friend, Manu ing from Anvil. 5 00:00:10,640 --> 00:00:11,810 Manu, thanks for joining today. 6 00:00:11,860 --> 00:00:13,660 Why don't you tell me a little bit about yourself, 7 00:00:13,720 --> 00:00:15,790 and I'd love to hear the elevator pitch for Anvil too. 8 00:00:15,840 --> 00:00:16,330 Awesome. 9 00:00:16,390 --> 00:00:18,190 Thanks for having me on the podcast today. 10 00:00:18,315 --> 00:00:18,470 Mike, it's. 11 00:00:19,050 --> 00:00:20,490 It's always a pleasure to do these. 12 00:00:20,830 --> 00:00:22,510 So a little bit about myself. 13 00:00:22,540 --> 00:00:25,020 Like Mike, Mike mentioned my name's Span ing. 14 00:00:25,420 --> 00:00:30,370 I am a founder, the founder of Anvil, and prior to that I worked as a 15 00:00:30,370 --> 00:00:31,780 software engineer for many years. 16 00:00:32,120 --> 00:00:37,030 So previously I had worked at companies like Dropbox Style Pad Flexport, and 17 00:00:37,030 --> 00:00:38,740 then also a small company called Loom. 18 00:00:39,470 --> 00:00:43,460 Not the video recording one that everybody knows about, but it, they did own loom.com 19 00:00:43,460 --> 00:00:45,200 and then that was acquired into Dropbox. 20 00:00:45,200 --> 00:00:47,910 So been in the startup space for quite some time. 21 00:00:48,240 --> 00:00:53,460 My background that kind of led me to Anvil was both personal and professional. 22 00:00:53,460 --> 00:00:57,730 I will say in my personal life I was just, dealing with adult stuff, 23 00:00:57,730 --> 00:01:01,780 like applying for a mortgage and getting infinite numbers of PDFs to 24 00:01:01,780 --> 00:01:03,790 fill out that I had to then send. 25 00:01:04,150 --> 00:01:08,630 Over email to bankers so that I could get a quote for a mortgage rate. 26 00:01:08,730 --> 00:01:09,870 So that was painful. 27 00:01:10,170 --> 00:01:13,720 And then in my professional life working at Dropbox, there was obviously 28 00:01:13,720 --> 00:01:15,880 a lot of documents at Flexport. 29 00:01:16,390 --> 00:01:20,620 What you don't realize is that most shipping is just generating 30 00:01:20,620 --> 00:01:22,870 documents along the way with some data that you have. 31 00:01:23,620 --> 00:01:27,250 Flexport system was essentially a system of record for the data 32 00:01:27,280 --> 00:01:30,300 and it would generate the correct documents as your shipment, cross 33 00:01:30,300 --> 00:01:32,520 borders and got onto different ships. 34 00:01:32,640 --> 00:01:37,960 And every time there was a PDF that needed to be created somebody had to hand code 35 00:01:37,960 --> 00:01:42,760 a template of that PDF in HTML, and then we'd render it and then generate the pdf. 36 00:01:42,760 --> 00:01:45,280 So that's why we created Anvil. 37 00:01:45,320 --> 00:01:47,210 Anvil is, frankly speaking the. 38 00:01:47,630 --> 00:01:52,310 Easiest way to build software for documents, whether that's generating PDFs, 39 00:01:52,400 --> 00:01:57,800 filling out PDFs, collecting e-signatures, or the entire document process where you 40 00:01:57,800 --> 00:02:01,310 gotta collect information, generate a certain set of documents based on some 41 00:02:01,310 --> 00:02:03,410 rules, and send it through E-signature. 42 00:02:03,860 --> 00:02:07,195 And the nice thing is that Anvil's completely embeddable white labeled 43 00:02:07,460 --> 00:02:10,850 and has a very robust API, which is why we're on the show today. 44 00:02:11,030 --> 00:02:11,510 right on. 45 00:02:11,570 --> 00:02:12,920 Lots to get into there. 46 00:02:13,000 --> 00:02:13,960 Really interesting. 47 00:02:13,960 --> 00:02:15,160 I know Flexport does a lot of 48 00:02:15,160 --> 00:02:16,300 Wild logistics stuff. 49 00:02:16,300 --> 00:02:18,520 I guess it totally makes sense that you're generating PDFs left 50 00:02:18,520 --> 00:02:21,160 and right 'cause there's bills of record the whole way through for 51 00:02:21,160 --> 00:02:22,420 shipping that are really important. 52 00:02:22,654 --> 00:02:23,374 That's fascinating. 53 00:02:23,374 --> 00:02:24,094 So you were, 54 00:02:24,344 --> 00:02:27,074 I guess in, in the world before Anvil existed, 55 00:02:27,264 --> 00:02:28,854 I'd imagine you felt the pain of 56 00:02:28,904 --> 00:02:30,254 Those PDFs in one way or the other. 57 00:02:30,554 --> 00:02:31,934 What was the thing that made you go 58 00:02:31,934 --> 00:02:35,264 Hey, like I'm a smart engineer, I should really dive into trying 59 00:02:35,264 --> 00:02:36,134 to make this a better thing. 60 00:02:36,690 --> 00:02:38,640 Smart engineer might be an overstatement. 61 00:02:38,640 --> 00:02:39,390 I was an engineer. 62 00:02:40,180 --> 00:02:40,840 Fair enough. 63 00:02:40,840 --> 00:02:41,260 Yeah. 64 00:02:41,280 --> 00:02:44,040 I've met very many engineers that are much smarter than I am. 65 00:02:44,070 --> 00:02:44,760 Let's put it that way. 66 00:02:45,240 --> 00:02:49,680 I think it was just, I knew I wanted to work on my own startup, 67 00:02:49,680 --> 00:02:51,060 my own project at some point. 68 00:02:51,060 --> 00:02:52,495 So that was always in the cards for me. 69 00:02:53,275 --> 00:02:57,595 I had been a part of many startups before starting Anvil, right? 70 00:02:57,595 --> 00:03:01,835 Like Dialpad, I was like number seven Lu I was number seven as well. 71 00:03:01,835 --> 00:03:04,655 And then Flexport and Dropbox was there during the growth stage 72 00:03:04,655 --> 00:03:06,335 is when it was like 500 people. 73 00:03:06,335 --> 00:03:08,225 So gotta see that spectrum. 74 00:03:09,020 --> 00:03:09,310 Sure. 75 00:03:09,995 --> 00:03:11,455 so that's maybe the starting point. 76 00:03:11,455 --> 00:03:16,195 And then I was really looking for a problem space that I felt was both. 77 00:03:16,630 --> 00:03:21,160 Large enough and painful enough, and also impactful enough that really 78 00:03:21,160 --> 00:03:25,010 would be something I'd be willing to commit to for the next, five, 79 00:03:25,010 --> 00:03:27,440 10, however many years to tackle. 80 00:03:27,440 --> 00:03:30,980 So I think those are probably the two driving forces. 81 00:03:31,080 --> 00:03:35,880 And then it was just really happenstance, like I was going through a period of my 82 00:03:35,880 --> 00:03:40,620 life when I was starting a company, where I just had a lot of documents to do. 83 00:03:40,650 --> 00:03:42,900 Like I was doing all the life stuff that you don't do when you're 84 00:03:42,900 --> 00:03:44,500 working, 'cause you put it off. 85 00:03:44,600 --> 00:03:47,750 And so it just was like an aha moment where everything clicked together. 86 00:03:47,750 --> 00:03:50,510 I think hindsight's 2020 always, but 87 00:03:50,560 --> 00:03:50,680 Sure. 88 00:03:50,725 --> 00:03:50,755 I. 89 00:03:51,238 --> 00:03:53,518 In my experience, I've been lucky to meet a lot of founders 90 00:03:53,518 --> 00:03:55,108 by way of podcast interview and, 91 00:03:55,388 --> 00:03:55,928 having friends 92 00:03:55,928 --> 00:03:56,768 With similar, 93 00:03:56,868 --> 00:03:59,118 Mental states to my own, who found themselves in 94 00:03:59,118 --> 00:04:00,378 The founder seat one way or the other. 95 00:04:00,378 --> 00:04:00,558 And 96 00:04:00,768 --> 00:04:01,248 There are a 97 00:04:01,298 --> 00:04:03,188 Few buckets that founders tend to fall into. 98 00:04:03,188 --> 00:04:03,548 And, 99 00:04:03,858 --> 00:04:05,928 There's always something interesting talking to people who've like really 100 00:04:05,928 --> 00:04:09,198 felt the problem that they're trying to solve before, before solving it 101 00:04:09,198 --> 00:04:12,378 because it's so much more deeply. 102 00:04:12,498 --> 00:04:12,738 Like 103 00:04:12,738 --> 00:04:16,188 You feel it to your core, but you know when it's success because 104 00:04:16,188 --> 00:04:17,538 past you would've loved this, 105 00:04:17,748 --> 00:04:18,348 To an extent. 106 00:04:18,698 --> 00:04:18,878 Yeah. 107 00:04:18,878 --> 00:04:19,628 That's really cool. 108 00:04:19,628 --> 00:04:19,898 I, 109 00:04:19,948 --> 00:04:23,248 I'd imagine a lot of the folks listening to this, I'd imagine virtually everyone 110 00:04:23,248 --> 00:04:24,988 on Earth has touched A PDF in some way. 111 00:04:25,288 --> 00:04:28,918 I'd imagine some of the people listening to this have tried to do what you 112 00:04:28,918 --> 00:04:34,318 described before with PDF templates or pulling structured data from a document or 113 00:04:34,528 --> 00:04:36,208 putting structured data into a document. 114 00:04:36,258 --> 00:04:36,948 my experience 115 00:04:36,948 --> 00:04:37,968 Is somewhat limited. 116 00:04:37,998 --> 00:04:38,928 But I can tell you that, 117 00:04:38,978 --> 00:04:41,498 I still have nightmares about trying to deal with some of the 118 00:04:41,521 --> 00:04:42,871 Biggest company in the world, like 119 00:04:42,871 --> 00:04:45,001 Their APIs and tools for building PDFs. 120 00:04:45,351 --> 00:04:45,741 Tell me about 121 00:04:45,741 --> 00:04:46,731 What the process is like. 122 00:04:46,731 --> 00:04:48,171 What's the hello world for someone, 123 00:04:48,201 --> 00:04:49,281 Onboarding with Anv look like 124 00:04:49,331 --> 00:04:49,781 Yeah. 125 00:04:49,781 --> 00:04:52,601 So typically the first thing you do is just sign up. 126 00:04:53,021 --> 00:04:56,531 You go grab a PDF form that you're trying to automate. 127 00:04:56,531 --> 00:04:58,451 So typically it's a blank form, right? 128 00:04:58,501 --> 00:04:59,856 We actually ask you to upload blank forms. 129 00:05:00,631 --> 00:05:02,911 And you upload it into Anvil. 130 00:05:03,121 --> 00:05:07,771 Anvil will automatically run through the PDF and use computer vision 131 00:05:07,771 --> 00:05:09,451 to find all the blanks for you. 132 00:05:10,021 --> 00:05:10,981 Does a pretty good job. 133 00:05:10,981 --> 00:05:14,091 I'll say not perfect, but does a very good job. 134 00:05:14,421 --> 00:05:18,506 And then we'll run it through an LLM to label and tag all the fields for you. 135 00:05:18,506 --> 00:05:20,096 So it'll do things like. 136 00:05:20,681 --> 00:05:22,901 Give each field a readable name. 137 00:05:22,901 --> 00:05:25,131 It'll give each field a specific type. 138 00:05:25,131 --> 00:05:29,451 So we have address types, social security number types, short text, long text, 139 00:05:29,451 --> 00:05:30,891 all that kind of stuff, check boxes. 140 00:05:31,451 --> 00:05:34,421 And it will also associate the fields together. 141 00:05:34,421 --> 00:05:37,571 So address fields typically are compound fields. 142 00:05:37,631 --> 00:05:40,221 There's line one, line two, city, state, zip. 143 00:05:41,621 --> 00:05:45,191 Anvil will associate those together so that when you go to 144 00:05:45,191 --> 00:05:48,626 populate this PDF, it's like a data structure that makes sense. 145 00:05:48,626 --> 00:05:51,776 It's not just a bunch of fields that are not interrelated, 146 00:05:51,776 --> 00:05:52,826 even though they should be. 147 00:05:53,196 --> 00:05:54,356 So that's pretty much it. 148 00:05:54,406 --> 00:05:58,876 If you have a schema already that you're trying to map to A PDF, you can pass 149 00:05:58,876 --> 00:06:03,256 along the keys and we will also try to associate your schema to specific 150 00:06:03,256 --> 00:06:04,996 fields on the PDF automatically. 151 00:06:05,536 --> 00:06:09,676 That just makes it easy for you to normalize data across multiple documents. 152 00:06:10,046 --> 00:06:12,866 And then you make an API request. 153 00:06:12,866 --> 00:06:16,196 We automatically generate a restful endpoint for you to fill this out. 154 00:06:16,556 --> 00:06:21,536 There's a payload of data that gets auto generated and you just make a request 155 00:06:21,536 --> 00:06:23,096 to that endpoint with your API key. 156 00:06:23,096 --> 00:06:25,046 And you should receive a filled out PDF. 157 00:06:25,640 --> 00:06:26,210 I'm a little, 158 00:06:26,290 --> 00:06:27,280 Gobsmacked right now. 159 00:06:27,280 --> 00:06:29,380 Having had tried to build something on, 160 00:06:29,430 --> 00:06:30,180 DocuSign, 161 00:06:30,430 --> 00:06:31,930 APIs in particular, like 162 00:06:32,090 --> 00:06:35,870 The complexity that large companies can inject into a process like 163 00:06:35,870 --> 00:06:37,610 this can be pretty mind numbing. 164 00:06:37,950 --> 00:06:39,060 And it's hard to see when you're that big. 165 00:06:39,060 --> 00:06:41,400 I've worked for great big corporations, I've worked for 166 00:06:41,400 --> 00:06:42,870 n equals one corporations, and, 167 00:06:43,150 --> 00:06:44,560 No fault in either direction. 168 00:06:45,120 --> 00:06:48,540 But like that, what you've just described sounds like the dream I needed, 169 00:06:48,570 --> 00:06:50,040 When I was working on this the last time. 170 00:06:50,330 --> 00:06:50,480 And 171 00:06:50,480 --> 00:06:52,040 It's a really interesting tool. 172 00:06:52,070 --> 00:06:55,310 Like that is something that I think fundamentally, anyone whose business 173 00:06:55,310 --> 00:06:58,430 has some overlap into the real world where a printer might be involved, 174 00:06:58,480 --> 00:06:59,230 Can use this 175 00:06:59,230 --> 00:07:00,160 And has some value for it. 176 00:07:00,300 --> 00:07:02,190 I'll actually maybe give you an example from my world. 177 00:07:02,200 --> 00:07:04,570 Outside of APIs, we won't hate or run a startup called Craftwork. 178 00:07:04,570 --> 00:07:07,300 We're a vertically integrated home painting company, and one of the things 179 00:07:07,300 --> 00:07:11,920 we deal with a lot is we get invoices and receipts from a paint supplier. 180 00:07:11,920 --> 00:07:13,030 So a paint store. 181 00:07:13,030 --> 00:07:14,770 We go and order paint for a project. 182 00:07:14,770 --> 00:07:15,130 It'll have 183 00:07:15,130 --> 00:07:15,940 The associated. 184 00:07:16,225 --> 00:07:17,275 Project that it's with. 185 00:07:17,275 --> 00:07:18,595 So it might be like Mike's house, 186 00:07:18,645 --> 00:07:20,955 And here's the 10 gallons of paint we ordered and the four paint 187 00:07:20,955 --> 00:07:23,835 brushes and the primer and all the other stuff that's needed there. 188 00:07:24,055 --> 00:07:26,785 Those per company are generally pretty standardized. 189 00:07:27,025 --> 00:07:31,135 However, that's like another form of oddly unstructured data in that 190 00:07:31,135 --> 00:07:32,185 it's an invoice that can have, 191 00:07:32,255 --> 00:07:33,995 One or 10 or 50 rows. 192 00:07:34,215 --> 00:07:36,315 And I don't necessarily have, let's call it like 193 00:07:36,325 --> 00:07:37,705 A blank version of that. 194 00:07:38,065 --> 00:07:39,385 In, in your world, is that 195 00:07:39,385 --> 00:07:39,775 Should I go 196 00:07:39,775 --> 00:07:43,135 And find a way to regenerate that sort of template with a blank? 197 00:07:43,325 --> 00:07:45,665 Is there a story for that with Anvil as well? 198 00:07:45,665 --> 00:07:46,445 Is that something that, 199 00:07:46,545 --> 00:07:47,865 Is a use case you're interested in? 200 00:07:47,865 --> 00:07:48,675 Yeah, absolutely. 201 00:07:48,675 --> 00:07:52,355 I think when it comes to extracting data out of PDFs, we 202 00:07:52,355 --> 00:07:53,645 don't actually do that today. 203 00:07:53,955 --> 00:07:57,555 Our thesis has always been, and I said we don't actually do that today. 204 00:07:57,555 --> 00:07:59,595 Actually, I don't know if we'll ever do it, let's be honest. 205 00:07:59,785 --> 00:08:03,205 Our thesis has always been that we already live in a world 206 00:08:03,205 --> 00:08:04,375 of a lot of structured data. 207 00:08:04,825 --> 00:08:08,215 And instead of destructuring it and putting it into a PDF, and then 208 00:08:08,215 --> 00:08:12,990 having somebody pull that unstructured data out of a PDF with OCR and lLM 209 00:08:12,990 --> 00:08:16,320 text extraction methods it should just be structured the whole time. 210 00:08:16,320 --> 00:08:19,260 Like the PDF becomes an artifact of the process. 211 00:08:19,420 --> 00:08:24,710 And so one way that I would think about your problem in this ideal future is 212 00:08:25,210 --> 00:08:32,540 when an invoice is generated as a PDF the actual data that went into generating 213 00:08:32,540 --> 00:08:35,720 the PDF should be attached in the PDF as metadata that is structured. 214 00:08:36,115 --> 00:08:39,625 And so if you're like receiving the PDF instead of parsing it out with computer 215 00:08:39,625 --> 00:08:43,335 vision and whatnot and overcomplicating it with this Rube Goldberg machine that 216 00:08:43,335 --> 00:08:47,060 you've created you just go read the data out and parse it into your system. 217 00:08:47,060 --> 00:08:50,630 And so this is actually like a hack week project that I did, which was I really 218 00:08:50,630 --> 00:08:53,810 wanted to take that structured data. 219 00:08:54,230 --> 00:08:56,450 Concept to the next level with PDFs. 220 00:08:56,550 --> 00:09:00,260 And Anvil does a lot with collecting data. 221 00:09:00,260 --> 00:09:02,000 We have a workflow tool, we call it. 222 00:09:02,090 --> 00:09:04,130 It collects data on the front end with a web form. 223 00:09:04,250 --> 00:09:05,060 It's dynamic. 224 00:09:05,120 --> 00:09:08,480 There's lots of logic, and then it generates the corresponding set 225 00:09:08,480 --> 00:09:10,070 of documents that you need, right? 226 00:09:10,910 --> 00:09:14,840 But really what I want to do is embed all that data as metadata 227 00:09:14,840 --> 00:09:19,950 alongside as the PDF, so that anybody could go pull it out if they knew 228 00:09:19,950 --> 00:09:21,960 it was a anvil generated document. 229 00:09:22,410 --> 00:09:25,680 Or hopefully this becomes like an open standard, and then PDFs 230 00:09:25,680 --> 00:09:27,210 literally become packets, right? 231 00:09:27,210 --> 00:09:29,940 It's like it's, there's like the internet packets and then there's 232 00:09:29,940 --> 00:09:33,120 like the packet that like the average, normal lay person would think of. 233 00:09:33,120 --> 00:09:36,510 It's like a packet of PDFs that then get passed around and the 234 00:09:36,510 --> 00:09:37,680 data can just be extracted. 235 00:09:38,290 --> 00:09:39,910 So that, that's our take on it. 236 00:09:39,990 --> 00:09:43,880 But we live in a messy world and there is a lot of a lot of. 237 00:09:44,480 --> 00:09:47,360 Unstructured data living in PDFs today, and I think there's lots of really 238 00:09:47,360 --> 00:09:49,160 great companies that are tackling it. 239 00:09:49,210 --> 00:09:54,530 From the, the legacy OCR companies and RPA companies to the newer crop of 'em 240 00:09:54,530 --> 00:09:56,570 that came along in this gen AI world. 241 00:09:56,925 --> 00:09:59,385 The idea that A PDF might have structured data embedded in 242 00:09:59,385 --> 00:10:00,555 it right now makes me wanna go 243 00:10:00,605 --> 00:10:01,835 Dive into my hex editor 244 00:10:01,835 --> 00:10:03,545 And figure out what I've been missing this whole time. 245 00:10:03,545 --> 00:10:05,165 Like genuinely, that's never crossed my mind 246 00:10:05,165 --> 00:10:06,035 That's even possible 247 00:10:06,050 --> 00:10:07,760 they, the funny thing is they don't, most of them do 248 00:10:07,895 --> 00:10:08,315 Yeah. 249 00:10:08,780 --> 00:10:13,070 it was really, actually, I, as far as I know, nobody, the, all of the 250 00:10:13,070 --> 00:10:18,140 metadata that gets embedded into the PDF is not the data that's on the PDF. 251 00:10:18,170 --> 00:10:20,870 It's it's oh, this person signed the document. 252 00:10:20,870 --> 00:10:22,220 It's that kind of information. 253 00:10:22,550 --> 00:10:24,380 But why not take it a step further? 254 00:10:24,380 --> 00:10:27,195 I don't understand why nobody's really done it that way, but 255 00:10:27,250 --> 00:10:27,460 That's 256 00:10:27,460 --> 00:10:29,290 No different than putting an image in a document. 257 00:10:29,450 --> 00:10:29,900 That's crazy. 258 00:10:30,080 --> 00:10:30,920 That's such a good idea. 259 00:10:30,950 --> 00:10:31,820 I really like that. 260 00:10:32,160 --> 00:10:33,450 You also mentioned documents, 261 00:10:33,480 --> 00:10:34,260 Signing as well. 262 00:10:34,260 --> 00:10:35,340 That's something that you serve. 263 00:10:35,670 --> 00:10:36,720 So it feels like you've got a 264 00:10:36,720 --> 00:10:37,710 Pretty complete picture for 265 00:10:37,750 --> 00:10:39,730 Populate this thing and then go ask for a signature. 266 00:10:39,760 --> 00:10:40,210 Is that right? 267 00:10:41,080 --> 00:10:41,950 Yeah, absolutely. 268 00:10:41,950 --> 00:10:46,170 We found that was like a very natural point for Anvil to move towards after 269 00:10:46,170 --> 00:10:47,790 building the first version of Anvil. 270 00:10:48,280 --> 00:10:51,165 We really think of paperwork as. 271 00:10:51,165 --> 00:10:55,875 Both a mechanism for data transport, like moving data between individuals and 272 00:10:55,875 --> 00:11:01,665 organizations, but also a mechanism for describing a business process, right? 273 00:11:01,665 --> 00:11:06,255 A lot of times a business operates on a set of documents and they 274 00:11:06,255 --> 00:11:09,615 have to be completed in a certain manner, in a certain order. 275 00:11:09,965 --> 00:11:16,375 And that essentially is a codification of what that company does and 276 00:11:16,425 --> 00:11:20,775 people leave, people come, but like the process stays the same. 277 00:11:21,435 --> 00:11:21,705 Yeah. 278 00:11:22,570 --> 00:11:24,310 It sounds like you've been through a few different phases. 279 00:11:24,310 --> 00:11:26,050 I don't know if you mentioned before, how old is the company? 280 00:11:27,005 --> 00:11:30,355 We're about seven and a half years so been around for a hot second. 281 00:11:30,625 --> 00:11:35,175 Back to the statement earlier about something I wanted to work on for five 10 282 00:11:35,175 --> 00:11:37,110 or maybe even longer in terms of years. 283 00:11:37,160 --> 00:11:37,670 Yeah. 284 00:11:37,670 --> 00:11:38,600 I'm curious then, like 285 00:11:38,600 --> 00:11:40,820 You've evolved over the past seven years. 286 00:11:41,030 --> 00:11:44,000 What did your first crop of customers look like and how did you find them? 287 00:11:44,000 --> 00:11:45,320 And then what do they look like now and 288 00:11:45,320 --> 00:11:46,310 How has that changed? 289 00:11:46,683 --> 00:11:46,863 yeah. 290 00:11:46,863 --> 00:11:51,088 Our first crop of customers we were really targeting non-technical people. 291 00:11:51,088 --> 00:11:53,758 The whole premise was, Hey, we're gonna create a set of tools. 292 00:11:54,198 --> 00:11:56,628 We actually, so we actually built the workflow tool first. 293 00:11:56,628 --> 00:11:58,758 A set of tools that where you can take documents. 294 00:11:58,758 --> 00:12:03,138 And upload it into Anvil and Anvil, then automatically create an online 295 00:12:03,138 --> 00:12:06,408 form for you that you can then share with people to collect the data. 296 00:12:06,408 --> 00:12:08,868 So the premise was TurboTax, anything. 297 00:12:09,258 --> 00:12:11,988 So our first customer was actually they're a great customer, 298 00:12:11,988 --> 00:12:13,038 still a customer of ours. 299 00:12:13,088 --> 00:12:13,898 First Ascent. 300 00:12:13,928 --> 00:12:15,788 They're now part of geo wealth. 301 00:12:15,818 --> 00:12:17,973 They are a giant, kinda like wealth management. 302 00:12:18,623 --> 00:12:20,153 Back office firms. 303 00:12:20,153 --> 00:12:23,873 So wealth advisors would choose them to manage all the documents and 304 00:12:23,873 --> 00:12:28,133 paperwork and they would handle the account opening documents, they would 305 00:12:28,133 --> 00:12:30,383 handle the transition, all that stuff. 306 00:12:30,433 --> 00:12:32,773 As well as like investment strategies and stuff like that. 307 00:12:32,773 --> 00:12:35,203 That's, that was our first customer. 308 00:12:35,203 --> 00:12:40,513 We effectively built Anvil around their use case. 309 00:12:40,563 --> 00:12:42,183 They were like a design partner, which was. 310 00:12:42,933 --> 00:12:47,748 Honestly a really lucky first experience, right? 311 00:12:47,828 --> 00:12:52,598 There was just so much about the way paperwork works and the way companies 312 00:12:52,598 --> 00:12:56,801 manage documents that, was so much more complicated than what we had anticipated. 313 00:12:56,981 --> 00:12:59,741 I think the most telling thing for me is like there's just a 314 00:12:59,741 --> 00:13:04,031 lot of tribal knowledge that lives in individual's heads that. 315 00:13:04,406 --> 00:13:08,226 If you're trying to codify it and translate it into a workflow you really 316 00:13:08,226 --> 00:13:10,296 need to like, pull it out of 'em. 317 00:13:10,386 --> 00:13:13,002 Oftentimes they don't even realize it's tribal knowledge and because 318 00:13:13,002 --> 00:13:16,782 it was such a complex use case, we actually built a pretty robust system. 319 00:13:17,292 --> 00:13:19,812 Fast forward a couple years, what we realized was. 320 00:13:19,812 --> 00:13:24,822 A lot of our friends were building companies in these legacy industries and 321 00:13:24,822 --> 00:13:28,302 they were running into this problem where they had to generate documents and they 322 00:13:28,302 --> 00:13:30,192 kept coming to us asking us for an API. 323 00:13:30,582 --> 00:13:34,532 So we just decided to productionize our own internal APIs, expose 324 00:13:34,532 --> 00:13:38,282 'em, make it easy for kinda like technology customers to use us. 325 00:13:38,282 --> 00:13:42,032 And nowadays the vast majority of our customers are actually 326 00:13:42,032 --> 00:13:45,902 like early to growth sage technology customers that find us. 327 00:13:46,232 --> 00:13:50,712 Read our APIs may maybe the lms the agents read our APIs and then 328 00:13:50,712 --> 00:13:53,082 implements us into their system. 329 00:13:53,132 --> 00:13:57,062 So yeah, we're pretty deeply embedded into organizations and 330 00:13:57,112 --> 00:14:00,762 we generally power like a pretty important stack and their business. 331 00:14:01,152 --> 00:14:01,242 Yeah. 332 00:14:01,287 --> 00:14:02,547 I will say, yeah, definitely. 333 00:14:02,547 --> 00:14:03,657 As someone, 334 00:14:03,657 --> 00:14:07,557 Building tools for a old industry that's a early to growth stage 335 00:14:07,557 --> 00:14:08,587 company, I think that's like. 336 00:14:09,222 --> 00:14:12,282 I can draw a lot of circles around things I need that are document driven right now. 337 00:14:12,282 --> 00:14:13,932 It's a really interesting problem to solve. 338 00:14:14,312 --> 00:14:17,072 And I can imagine many types of companies that would be into that. 339 00:14:17,472 --> 00:14:19,962 I'm curious what you can tell me about how the software is built. 340 00:14:19,962 --> 00:14:22,182 So from architecture to, 341 00:14:22,462 --> 00:14:23,992 the types of engineers that work for you. 342 00:14:23,992 --> 00:14:25,432 What are, how are you building this? 343 00:14:25,432 --> 00:14:25,612 And, 344 00:14:25,832 --> 00:14:27,002 I dunno, what does your team look like? 345 00:14:27,822 --> 00:14:28,152 Yeah. 346 00:14:28,152 --> 00:14:31,812 So from a technology stack perspective, it's a pretty standard stack. 347 00:14:31,812 --> 00:14:32,862 It's no js. 348 00:14:33,307 --> 00:14:36,487 GraphQL endpoints and react on the front end. 349 00:14:36,937 --> 00:14:43,437 We also have a fairly robust like PDF library that we it's an open 350 00:14:43,437 --> 00:14:47,757 source library that we've then wrapped in this nice web experience. 351 00:14:47,757 --> 00:14:50,757 So there's a lot, a fair amount of Java in our code as well. 352 00:14:50,837 --> 00:14:52,037 For better or worse, and. 353 00:14:53,132 --> 00:14:55,592 Yeah so that's the joke is always oh, it makes sense that the 354 00:14:55,592 --> 00:15:00,182 ultimate enterprise file format is also written in the ultimate 355 00:15:00,182 --> 00:15:02,312 enterprise language, which is Java. 356 00:15:02,322 --> 00:15:03,042 Of course. 357 00:15:03,142 --> 00:15:04,612 And so that's the stack. 358 00:15:04,712 --> 00:15:06,932 In terms of our team, we're a very lean team. 359 00:15:06,982 --> 00:15:10,272 We're a team of seven and it's about half engineering and. 360 00:15:10,942 --> 00:15:12,232 Half non-engineering. 361 00:15:12,572 --> 00:15:15,212 Everybody's fairly technical, I will say on our team. 362 00:15:15,262 --> 00:15:17,812 And I say half engineering 'cause I count myself as half an engineer, 363 00:15:17,812 --> 00:15:19,342 so that makes it exactly half. 364 00:15:19,622 --> 00:15:23,372 And yeah and so it's, we run pretty lean. 365 00:15:23,402 --> 00:15:27,662 The nice thing is our product is, generally speaking, very self-serve. 366 00:15:27,812 --> 00:15:32,552 We tried a heavy sales motion early on, but just given our ICP, which 367 00:15:32,552 --> 00:15:34,772 is engineers and product people. 368 00:15:36,122 --> 00:15:37,232 didn't want to get on sales calls. 369 00:15:37,262 --> 00:15:39,362 We just, they just wanted to have access to the docks. 370 00:15:39,362 --> 00:15:42,842 They wanted to have a dev sandbox. 371 00:15:42,842 --> 00:15:44,012 They wanted to kick the tires. 372 00:15:44,042 --> 00:15:47,282 And that's how a lot of our customers come to us. 373 00:15:47,322 --> 00:15:53,712 They start paying us 5 cents, 10 cents, a hundred dollars, $2,000 a 374 00:15:53,712 --> 00:15:56,472 month, and then it just becomes like a real, it becomes a real account. 375 00:15:56,532 --> 00:15:56,862 Yeah. 376 00:15:57,522 --> 00:15:57,912 Yeah, 377 00:15:57,912 --> 00:15:58,122 I, 378 00:15:58,312 --> 00:16:00,832 Am guilty of diving into your docs and looking at things, 379 00:16:00,862 --> 00:16:02,602 as I've been like preparing for this interview. 380 00:16:02,602 --> 00:16:02,752 And, 381 00:16:03,132 --> 00:16:04,542 A few things stood out to me that 382 00:16:04,542 --> 00:16:08,802 are very bright green signals for me as a developer is, I'll start by saying that 383 00:16:08,862 --> 00:16:11,622 if you're listening to this podcast, you should go check out Anvil's docs pages. 384 00:16:11,652 --> 00:16:13,002 'cause they're so well organized. 385 00:16:13,002 --> 00:16:14,562 They're beautiful, they're like in 386 00:16:14,752 --> 00:16:15,292 Thought out. 387 00:16:15,292 --> 00:16:16,582 They have nice API reference. 388 00:16:16,582 --> 00:16:18,832 It's like whoever worked on those, give them a high five for me. 389 00:16:18,832 --> 00:16:19,222 That's, 390 00:16:19,552 --> 00:16:23,662 What I would say is do that and then think of any other PDF, like API you can 391 00:16:23,662 --> 00:16:25,162 think of and go look at the difference. 392 00:16:25,162 --> 00:16:25,312 And 393 00:16:25,312 --> 00:16:27,172 As an engineer, I feel like you can feel the difference. 394 00:16:27,222 --> 00:16:30,012 If you're selling to engineers, in my eyes, that sells itself. 395 00:16:30,432 --> 00:16:33,252 I've, I'm also like really happy to see that, 396 00:16:33,302 --> 00:16:34,652 Your pricing is pretty transparent. 397 00:16:34,812 --> 00:16:36,522 Do you wanna talk a little bit about the pricing strategy? 398 00:16:37,352 --> 00:16:38,132 Yeah, absolutely. 399 00:16:38,182 --> 00:16:40,072 It is pretty much pay as you go. 400 00:16:40,352 --> 00:16:42,332 It's 10 cents A PDF generation. 401 00:16:42,332 --> 00:16:45,302 It's a dollar 50 for a completed signature packet. 402 00:16:45,302 --> 00:16:46,487 And I say completed because. 403 00:16:47,487 --> 00:16:51,717 All of the other, or not all, but 99% of the other e-sign providers, 404 00:16:51,717 --> 00:16:55,737 they charge you on send regardless of if the signatures were completed. 405 00:16:56,247 --> 00:16:58,627 And so we really want to align value there. 406 00:16:58,627 --> 00:17:01,417 And then it's a dollar for a workflow submission. 407 00:17:01,417 --> 00:17:05,797 So a workflow is a web form that can be pretty complex or multiple web 408 00:17:05,797 --> 00:17:07,987 forms that then populate multiple PDFs. 409 00:17:08,087 --> 00:17:12,247 We do also have some kind of like platform fee type fee pricing. 410 00:17:12,862 --> 00:17:16,582 Specifically if you want like white labeling or some high, more advanced 411 00:17:16,582 --> 00:17:20,962 features, those are a monthly platform fee or an annual platform fee. 412 00:17:21,482 --> 00:17:24,002 And then enterprise is similar. 413 00:17:24,002 --> 00:17:25,232 It's a larger platform fee. 414 00:17:25,232 --> 00:17:30,062 You have more advanced features, you get more support, and then of course, as usage 415 00:17:30,062 --> 00:17:32,522 scales, you're able to buy discounted. 416 00:17:32,897 --> 00:17:33,437 Usage. 417 00:17:33,807 --> 00:17:37,357 So overall we want to scale with you. 418 00:17:37,357 --> 00:17:41,317 We wanna make it easy for you to start and not pay a bunch up front, but 419 00:17:41,317 --> 00:17:44,347 if you do start using us a lot, we don't want to be nickel and dimming 420 00:17:44,347 --> 00:17:45,947 you at retail prices, essentially. 421 00:17:46,382 --> 00:17:47,612 What I like about that is, 422 00:17:47,642 --> 00:17:48,542 It's predictable. 423 00:17:48,542 --> 00:17:50,642 If you can look at that and understand how it fits into your 424 00:17:50,642 --> 00:17:52,052 world as a team, building something, 425 00:17:52,392 --> 00:17:53,622 But also accessible enough, 426 00:17:53,652 --> 00:17:54,942 If I want to test it and see if, 427 00:17:54,972 --> 00:17:56,682 My first a hundred documents, 428 00:17:56,682 --> 00:17:59,502 Makes sense and work and fits my team's use case, that's 429 00:17:59,682 --> 00:18:01,122 not a insurmountable thing, 430 00:18:01,172 --> 00:18:03,152 The first a hundred documents will actually be free. 431 00:18:03,182 --> 00:18:05,372 'cause when you plop on a credit card, we give you a bunch of 432 00:18:05,372 --> 00:18:07,162 free credit for production usage. 433 00:18:07,552 --> 00:18:10,152 So I believe it's don't quote me on this. 434 00:18:10,182 --> 00:18:13,062 I believe it's a thousand PDF generations and. 435 00:18:13,782 --> 00:18:17,742 Like 50 workflows or 50 signatures. 436 00:18:17,742 --> 00:18:19,542 I'll have to double check, but it is pretty generous. 437 00:18:19,542 --> 00:18:23,552 We do want you to I think we want you to build as a developer and use 438 00:18:23,552 --> 00:18:28,652 the dev keys and all that stuff, and that's pretty much the full experience. 439 00:18:28,652 --> 00:18:32,872 You get access to pretty much everything When you flip into production, we remove 440 00:18:32,872 --> 00:18:34,012 the watermarks and stuff like that. 441 00:18:34,012 --> 00:18:35,932 We do want to give you enough credit to actually. 442 00:18:36,292 --> 00:18:40,437 Run it into production in a meaningful way so that you have confidence that you 443 00:18:40,437 --> 00:18:41,967 can build your business on top of this. 444 00:18:42,712 --> 00:18:43,552 Yeah, I like that too. 445 00:18:43,552 --> 00:18:44,332 I walked right into 446 00:18:44,332 --> 00:18:45,412 Your free tier as well. 447 00:18:45,412 --> 00:18:45,652 I, 448 00:18:45,702 --> 00:18:46,152 Promise I 449 00:18:46,152 --> 00:18:47,472 Didn't try to do that, but 450 00:18:47,472 --> 00:18:48,072 That's true. 451 00:18:48,312 --> 00:18:50,322 And so from what I remember, looking through your docs, 452 00:18:50,372 --> 00:18:52,407 You are GraphQL and rest endpoint 453 00:18:52,407 --> 00:18:53,037 Driven, right? 454 00:18:53,037 --> 00:18:53,637 So there's like 455 00:18:53,827 --> 00:18:54,397 Our API, 456 00:18:54,737 --> 00:18:57,677 Homies listening to the podcast will be really interested in what that all, 457 00:18:57,867 --> 00:18:58,497 Looks like. 458 00:18:58,827 --> 00:18:59,457 Do you have, 459 00:18:59,487 --> 00:18:59,937 is it, 460 00:18:59,967 --> 00:19:01,467 Primarily like rest endpoints 461 00:19:01,467 --> 00:19:02,172 In GraphQL? 462 00:19:02,172 --> 00:19:02,562 Do you have, 463 00:19:02,612 --> 00:19:04,262 Client SDKs in, written in, 464 00:19:04,312 --> 00:19:05,542 Java, whatever c 465 00:19:05,942 --> 00:19:07,202 Node languages Or is it, 466 00:19:07,252 --> 00:19:08,332 Or documentation around 467 00:19:08,492 --> 00:19:10,247 Send a request that looks like this and this is what will happen. 468 00:19:11,602 --> 00:19:12,052 Yeah. 469 00:19:12,107 --> 00:19:15,947 So we do have client sks that we've written and we publish. 470 00:19:16,047 --> 00:19:21,477 They are, I believe there's a JavaScript one, a C one, there's a Python one. 471 00:19:22,172 --> 00:19:25,532 And we generally try to be pretty good about maintaining them. 472 00:19:25,582 --> 00:19:29,182 Occasionally we do get some people writing in and then we dive in there and go fix 473 00:19:29,182 --> 00:19:33,502 it, and we do ask if somebody has a fix, please feel free to open up pull request. 474 00:19:33,902 --> 00:19:35,462 So we do have a lot of pre-written libraries. 475 00:19:36,062 --> 00:19:38,882 They're mostly just like client wrappers. 476 00:19:38,932 --> 00:19:44,342 Our API is pretty hopefully pretty, pretty simple to understand and integrate with. 477 00:19:44,442 --> 00:19:46,332 And so I think that. 478 00:19:46,432 --> 00:19:49,242 If you're just making web requests it's pretty easy. 479 00:19:49,292 --> 00:19:53,912 The client libraries do nice things like handle multi-part uploads for 480 00:19:53,912 --> 00:19:58,262 you if you're uploading a large PDF to be templatized or handle, 481 00:19:58,262 --> 00:20:01,022 like automatically handle rate limiting four, two nines for you. 482 00:20:01,022 --> 00:20:04,142 And using your encryption keys and stuff like that. 483 00:20:04,142 --> 00:20:08,552 So it's really just like nice helper functions to, to get you started faster. 484 00:20:09,082 --> 00:20:11,092 One of the other things I wanted to mention too is you have quite 485 00:20:11,092 --> 00:20:14,242 A broad list of like open source things published on Anil's, 486 00:20:14,292 --> 00:20:15,612 GitHub, which is really cool to see. 487 00:20:15,762 --> 00:20:16,662 And also nice that 488 00:20:16,662 --> 00:20:17,052 You're, 489 00:20:17,152 --> 00:20:19,852 Open to and willing to like, accept community driven, 490 00:20:19,902 --> 00:20:20,952 Changes and whatnot. 491 00:20:21,232 --> 00:20:21,802 Especially, 492 00:20:21,802 --> 00:20:23,872 Keeping up with lots of languages can be a challenge and, 493 00:20:24,152 --> 00:20:26,702 getting the, like developer flavored change 494 00:20:26,702 --> 00:20:27,482 Is often. 495 00:20:27,642 --> 00:20:28,782 Really useful, especially, 496 00:20:28,822 --> 00:20:31,387 If it's like not your specific native language, 497 00:20:31,487 --> 00:20:32,507 Yeah, absolutely. 498 00:20:32,507 --> 00:20:32,987 Absolutely. 499 00:20:32,987 --> 00:20:34,847 And I'll say, actually just a quick call out. 500 00:20:34,847 --> 00:20:41,357 If anybody out there is looking for a documentation reference docs generation 501 00:20:41,357 --> 00:20:48,767 system for GraphQL check out spectacle, that's S-P-E-C-T-A-Q-L that is the 502 00:20:48,767 --> 00:20:53,207 system we use and we op, we open source for generating our reference docs for 503 00:20:53,412 --> 00:20:53,522 very cool. 504 00:20:54,367 --> 00:20:58,117 It was a fork from a previous project that kind of got abandoned and we updated it. 505 00:20:58,117 --> 00:20:59,017 And yeah. 506 00:20:59,077 --> 00:21:02,147 Now whenever we push a new endpoint or change something it just 507 00:21:02,147 --> 00:21:06,287 automatically compiles and creates a new reference stocks for our endpoints. 508 00:21:07,127 --> 00:21:07,937 I will make sure 509 00:21:07,937 --> 00:21:09,047 That shows up in the show notes. 510 00:21:09,047 --> 00:21:11,147 So if you're listening to this, you can scroll down to the description of 511 00:21:11,147 --> 00:21:12,677 your podcast and get right to that too. 512 00:21:13,117 --> 00:21:13,717 Can you tell me like 513 00:21:13,717 --> 00:21:15,157 Where your team is headed? 514 00:21:15,157 --> 00:21:16,717 What are the kinds of problems you're looking at next? 515 00:21:17,857 --> 00:21:18,247 Oh man. 516 00:21:18,407 --> 00:21:18,947 So many. 517 00:21:18,997 --> 00:21:22,807 I think there's one to start maybe revisiting the pricing question earlier. 518 00:21:23,107 --> 00:21:25,327 One thing we did last year was. 519 00:21:25,712 --> 00:21:29,852 Make any usage of Anvil from our online experience free. 520 00:21:30,132 --> 00:21:32,292 The goal is really to make Anvil accessible to 521 00:21:32,502 --> 00:21:33,852 non-technical people as well. 522 00:21:33,852 --> 00:21:38,382 So your ops team can send out a signature packet and have that be free. 523 00:21:38,682 --> 00:21:41,952 And so that kind of plays into what we're thinking about in 2026, 524 00:21:42,162 --> 00:21:45,762 which is we want to bring anvil to places that people do work. 525 00:21:46,092 --> 00:21:48,642 And this is I think, a pretty strong counter position 526 00:21:48,642 --> 00:21:50,172 against other e-sign providers. 527 00:21:50,202 --> 00:21:52,872 Other e-sign providers think of themselves as the platform and then 528 00:21:52,872 --> 00:21:56,682 say, Hey, come to us and we will be your one place to manage documents. 529 00:21:56,682 --> 00:22:00,022 And the fact of the matter is like nobody's working out of DocuSign. 530 00:22:00,112 --> 00:22:02,512 They're working out of Salesforce, they're working out of HubSpot. 531 00:22:03,142 --> 00:22:07,882 And so what Anvil's trying to do is make it so they can connect 532 00:22:07,882 --> 00:22:09,652 Anvil to your Salesforce account. 533 00:22:10,747 --> 00:22:15,387 And then from within Salesforce, let's say you can just say, generate us a sales 534 00:22:15,387 --> 00:22:17,127 contract and automatically generates it. 535 00:22:17,277 --> 00:22:20,457 You review it with our embedded components all within Salesforce, 536 00:22:20,517 --> 00:22:21,747 and then send out for signature. 537 00:22:21,747 --> 00:22:26,067 So we wanna do this for Salesforce, HubSpot, all of these like touch points 538 00:22:26,067 --> 00:22:28,497 where teams generally get work done. 539 00:22:29,492 --> 00:22:32,552 And a big one of, for the, for that in the, forward looking 540 00:22:32,552 --> 00:22:34,442 is really MCP UI apps, right? 541 00:22:34,492 --> 00:22:35,392 Yeah, sure. 542 00:22:35,562 --> 00:22:39,412 chat, GPT, Claude Cowork where somebody can say, I need to get this 543 00:22:39,412 --> 00:22:44,302 document signed, upload an anvil's system automatically tags and labels. 544 00:22:44,302 --> 00:22:45,832 And they say here's a payload of data. 545 00:22:45,862 --> 00:22:46,552 Go fill it out. 546 00:22:46,552 --> 00:22:47,722 Send it out for signature. 547 00:22:47,822 --> 00:22:48,932 That's one aspect. 548 00:22:49,032 --> 00:22:52,602 The flip side of that is how do we prevent agents from just randomly 549 00:22:52,602 --> 00:22:55,272 signing a bunch of stuff like we need to know it's a human. 550 00:22:55,572 --> 00:22:59,952 So we got some pretty fun stuff fun projects up our sleeves to create 551 00:22:59,952 --> 00:23:05,172 a frictionless way to identify an individual as a human, not an agent. 552 00:23:05,682 --> 00:23:07,992 More on that to come, but that's something we're working on 553 00:23:07,992 --> 00:23:10,182 probably in, in Q1, Q2 this year. 554 00:23:10,692 --> 00:23:13,972 And I think we'll be the first ones to really tackle that problem beyond like a. 555 00:23:14,842 --> 00:23:18,022 Hey, set up a video camera, a video call and notarize and have 556 00:23:18,022 --> 00:23:21,042 a notary or do one of those like crazy face scan things, right? 557 00:23:21,042 --> 00:23:22,482 That nobody seems to like doing. 558 00:23:22,482 --> 00:23:24,462 So that's another one. 559 00:23:24,562 --> 00:23:29,462 And then the last project, which is the bid project is goes back to 560 00:23:29,462 --> 00:23:33,902 how thinks of ourselves as a data company, not so much a PDF company. 561 00:23:33,962 --> 00:23:36,752 Like the data, just the PDF is an artifact and it just happens 562 00:23:36,752 --> 00:23:37,982 to be how the world operates. 563 00:23:38,787 --> 00:23:43,887 But one thing we do today is again, you can upload your keys when 564 00:23:43,887 --> 00:23:48,177 you're tagging A PDF, and we will automatically match your key schema 565 00:23:48,327 --> 00:23:50,697 to the schema of the document, right? 566 00:23:50,697 --> 00:23:51,777 So that makes it easy for you. 567 00:23:51,777 --> 00:23:55,137 Then create data payloads that are easy for you to manipulate in your 568 00:23:55,137 --> 00:23:57,477 code and have that rendered onto A PDF. 569 00:23:58,807 --> 00:23:59,047 With. 570 00:24:00,112 --> 00:24:04,072 new project, what we're trying to do is go beyond mapping a schema to A PDF. 571 00:24:04,072 --> 00:24:08,482 Imagine a world where you could grab a payload of data in any schema 572 00:24:08,782 --> 00:24:10,372 and tell us the target schema. 573 00:24:10,642 --> 00:24:13,672 And an and anvil just automatically translates that 574 00:24:13,672 --> 00:24:15,742 data into your target schema. 575 00:24:15,742 --> 00:24:17,182 So you can use it in whatever system. 576 00:24:17,182 --> 00:24:20,542 It could be a PDF, it could be another application database, 577 00:24:20,542 --> 00:24:22,462 it could be your snowflake. 578 00:24:22,862 --> 00:24:25,802 So we already have this general concept in Anvil. 579 00:24:25,802 --> 00:24:27,302 It's just very PDF oriented. 580 00:24:27,842 --> 00:24:28,132 Sure. 581 00:24:28,967 --> 00:24:32,807 The goal is to build it out so that you can effectively install anvil 582 00:24:32,897 --> 00:24:38,197 in front of your application or, wherever and just use it as a pass 583 00:24:38,197 --> 00:24:42,987 through for transforming data on the fly to be consumed in another system. 584 00:24:43,057 --> 00:24:45,757 So that, that's really what we're looking to, to work on 585 00:24:45,767 --> 00:24:47,177 in this year and in the future. 586 00:24:48,137 --> 00:24:49,667 It's a really cool use of the tools and 587 00:24:49,667 --> 00:24:51,347 Expanding your borders a little bit. 588 00:24:51,597 --> 00:24:53,577 And becoming the translator between, 589 00:24:53,857 --> 00:24:54,637 Adjacent, 590 00:24:55,127 --> 00:24:57,287 Namespace is a really interesting idea. 591 00:24:57,627 --> 00:25:00,267 I have spent a lot of my career, and I'm sure a lot of people listening 592 00:25:00,267 --> 00:25:02,367 to this have spent a lot of their career doing exactly that, where it's 593 00:25:02,367 --> 00:25:03,327 I don't know, we just get this. 594 00:25:03,552 --> 00:25:06,792 Garbage format from some COBOL program from 700 years ago, and 595 00:25:06,792 --> 00:25:08,472 We need to convert it into this new thing. 596 00:25:08,652 --> 00:25:09,702 And so I'm gonna spend the next, 597 00:25:09,912 --> 00:25:13,302 Nine weeks writing a, an intensive like JSON schema thing 598 00:25:13,302 --> 00:25:14,592 to translate things around. 599 00:25:14,592 --> 00:25:14,772 And, 600 00:25:15,092 --> 00:25:18,122 A lot of the world we're living in now makes that easier and more interesting, 601 00:25:18,192 --> 00:25:19,062 And more capable for 602 00:25:19,062 --> 00:25:19,902 Everyone to do. 603 00:25:20,292 --> 00:25:21,012 That's super cool. 604 00:25:21,192 --> 00:25:21,612 I really like 605 00:25:21,612 --> 00:25:21,822 Yeah. 606 00:25:21,832 --> 00:25:24,112 Did I say the words AI yet, or the word AI yet? 607 00:25:25,202 --> 00:25:28,072 The part of the reason of doing this is actually, if you think about is actually 608 00:25:28,072 --> 00:25:31,992 really good at these types of matching problems where it's you have this thing 609 00:25:31,992 --> 00:25:33,162 here and you have this thing here. 610 00:25:33,162 --> 00:25:36,042 And like a human could look at it and be like, oh yeah, that makes sense. 611 00:25:36,042 --> 00:25:38,112 This field obviously meet matches this key. 612 00:25:38,542 --> 00:25:41,672 But to do that derma deterministically is quite difficult. 613 00:25:41,702 --> 00:25:45,272 You'll never run out of rules to have to implement, but to do 614 00:25:45,272 --> 00:25:49,022 that with ai and lms, it's a lot. 615 00:25:49,117 --> 00:25:49,777 Easier. 616 00:25:49,777 --> 00:25:54,217 And so now we actually have this tool that allows us to do that, and we think 617 00:25:54,217 --> 00:25:58,327 we have a unique data set that allows us to do that better than anybody else. 618 00:25:58,357 --> 00:26:02,967 And we can actually fine tune and tr, make these models work for this use case. 619 00:26:03,367 --> 00:26:07,547 That I think is actually a pretty good moat as we build up this feature. 620 00:26:07,607 --> 00:26:07,787 Yeah. 621 00:26:08,087 --> 00:26:08,627 Definitely. 622 00:26:08,687 --> 00:26:09,917 Yeah, that sounds really interesting. 623 00:26:10,077 --> 00:26:12,357 You'll have to send that my way as soon as it's available. 624 00:26:12,357 --> 00:26:13,377 I'd love to play around with it. 625 00:26:14,097 --> 00:26:15,057 I definitely will. 626 00:26:16,017 --> 00:26:17,457 Couple more small things for you. 627 00:26:17,507 --> 00:26:18,407 Is your team expanding? 628 00:26:18,407 --> 00:26:19,577 Are you looking to hire anytime soon? 629 00:26:20,447 --> 00:26:21,167 Yeah, absolutely. 630 00:26:21,167 --> 00:26:24,067 We are hiring specifically looking for engineers. 631 00:26:24,117 --> 00:26:30,627 We are planning on hiring two-ish, two to three-ish engineers, so if anybody's 632 00:26:30,627 --> 00:26:32,697 interested, please do check us out. 633 00:26:33,007 --> 00:26:36,487 I believe the jobs are posted on, found, also known as Angel List. 634 00:26:36,567 --> 00:26:43,727 Anybody has deep experience with AI ml and wants to be a full stack, just 635 00:26:43,727 --> 00:26:46,517 like a builder, I don't wanna say you're not just focused on AI and ml, 636 00:26:46,517 --> 00:26:49,787 but you can be building the entire application and be close to customers. 637 00:26:50,207 --> 00:26:51,337 Yeah, come check us out. 638 00:26:51,337 --> 00:26:53,582 We have some we have some open roles there. 639 00:26:54,122 --> 00:26:54,692 Very cool. 640 00:26:54,722 --> 00:26:55,082 Yeah, 641 00:26:55,087 --> 00:26:56,062 I will be sure to, 642 00:26:56,112 --> 00:26:57,612 Stick that link in the bottom here too. 643 00:26:58,052 --> 00:27:00,452 The last thing I wanted to ask you about is you host a podcast as well. 644 00:27:00,452 --> 00:27:01,262 Can you tell me about that? 645 00:27:01,262 --> 00:27:01,817 Gimme the pitch for your show. 646 00:27:03,182 --> 00:27:03,632 Yeah. 647 00:27:03,632 --> 00:27:04,202 Thanks Mike. 648 00:27:04,252 --> 00:27:10,002 We host a podcast called Build First Buy and we get product leaders founders 649 00:27:10,002 --> 00:27:14,262 on our show to talk about their decisions when they're building a 650 00:27:14,262 --> 00:27:15,912 new company, creating a new company. 651 00:27:16,357 --> 00:27:20,167 Their decision making around whether to build something or buy something 652 00:27:20,167 --> 00:27:21,917 and specifically around technology. 653 00:27:21,917 --> 00:27:26,237 And I think there is a strong bias to building in Silicon Valley. 654 00:27:26,267 --> 00:27:30,977 I think one guest that came on made a very good point that like people 655 00:27:30,977 --> 00:27:35,067 in Silicon Valley are rewarded for building but fundamentally there are 656 00:27:35,067 --> 00:27:36,537 just things that you probably shouldn't. 657 00:27:37,122 --> 00:27:39,882 Build if you're really trying to move fast and launch. 658 00:27:39,882 --> 00:27:41,202 That's the conversation we have there. 659 00:27:41,202 --> 00:27:45,702 We've had 26, 27 episodes now with some really great leaders. 660 00:27:46,132 --> 00:27:49,852 Many of them are people that I've been fortunate enough to work with in the past. 661 00:27:49,902 --> 00:27:51,582 And yeah, would love for people to check it out. 662 00:27:51,582 --> 00:27:55,752 It's bill verse buy look it up with Anvil in the Google search. 663 00:27:55,752 --> 00:27:57,642 Or ask Claude to look it up for you or something. 664 00:27:58,302 --> 00:27:58,482 Yeah. 665 00:27:58,487 --> 00:27:58,647 And. 666 00:27:59,427 --> 00:28:02,367 Hopefully you guys enjoyed listening to that podcast as well. 667 00:28:02,832 --> 00:28:03,642 Yeah, definitely. 668 00:28:03,642 --> 00:28:05,112 We'll make sure to link that through too. 669 00:28:05,212 --> 00:28:07,132 Build versus buy is the perpetual question, 670 00:28:07,182 --> 00:28:07,932 For all of us. 671 00:28:07,932 --> 00:28:08,082 And 672 00:28:08,302 --> 00:28:08,812 I agree with you. 673 00:28:08,812 --> 00:28:12,592 I think a lot of Silicon Valley rewards building when like money 674 00:28:12,592 --> 00:28:14,122 works to make things happen too, 675 00:28:14,167 --> 00:28:15,727 We see that now as well, right? 676 00:28:15,727 --> 00:28:18,337 Like the whole premise around like Clause is gonna build everything. 677 00:28:18,337 --> 00:28:21,247 You're just gonna prompt claw code and know SaaS is dead. 678 00:28:21,297 --> 00:28:22,377 Maybe we'll get there. 679 00:28:22,407 --> 00:28:23,427 Maybe that is the future. 680 00:28:23,427 --> 00:28:27,057 But I think as of the current state, Claude is great at building 681 00:28:27,057 --> 00:28:30,687 MVPs and actually, and great at editing existing code bases. 682 00:28:30,987 --> 00:28:33,237 I think that's what it's really good at and it is very powerful. 683 00:28:33,237 --> 00:28:38,847 I use it every day, but to productionize something to make it scalable and reliable 684 00:28:38,847 --> 00:28:43,197 and resilient and you honestly have to go through a couple fires and feel the pain. 685 00:28:43,827 --> 00:28:43,977 No. 686 00:28:43,977 --> 00:28:46,212 And then get a bunch of phone calls and then, yeah. 687 00:28:46,212 --> 00:28:49,182 So we, I think we've battle tested and battle hardened our 688 00:28:49,182 --> 00:28:50,652 system quite a bit at this point. 689 00:28:50,712 --> 00:28:51,942 You're speaking to my heart. 690 00:28:51,942 --> 00:28:54,042 I feel that deeply right now, without a doubt. 691 00:28:54,792 --> 00:28:55,272 Right on. 692 00:28:55,342 --> 00:28:56,812 Manding, thank you for joining. 693 00:28:56,842 --> 00:28:58,402 I really appreciate having you on the show. 694 00:28:58,512 --> 00:29:01,062 Before I let you go, where's the best place to find you online and 695 00:29:01,062 --> 00:29:02,382 how do people find Anvil Online? 696 00:29:03,267 --> 00:29:03,597 Yeah. 697 00:29:03,597 --> 00:29:05,457 Best place to find me online is LinkedIn. 698 00:29:05,517 --> 00:29:06,117 Mange it. 699 00:29:06,117 --> 00:29:09,807 Just plot or just like Google, me, my name's unique enough that I'm, I think 700 00:29:09,807 --> 00:29:11,757 I'm the first two pages of results. 701 00:29:11,947 --> 00:29:14,497 I most active on LinkedIn, I will say. 702 00:29:14,557 --> 00:29:17,617 And then Anvil is use anvil.com. 703 00:29:18,022 --> 00:29:19,552 We also have a YouTube channel. 704 00:29:19,582 --> 00:29:22,852 We have a TikTok and Instagram, which you might think is funny 705 00:29:22,852 --> 00:29:25,582 for a B2B SaaS company, but there's some funny videos there. 706 00:29:25,582 --> 00:29:27,712 I would say we, we work with a really great agency. 707 00:29:27,762 --> 00:29:28,782 So check those out. 708 00:29:28,782 --> 00:29:32,652 Hopefully they're entertaining and yeah, hope to hear from you. 709 00:29:32,712 --> 00:29:35,557 I'll maybe on LinkedIn or on one of our social. 710 00:29:36,887 --> 00:29:37,307 Right on. 711 00:29:37,307 --> 00:29:37,787 That sounds great. 712 00:29:37,787 --> 00:29:39,137 We'll make sure to tag you there when this, 713 00:29:39,137 --> 00:29:39,947 Goes live too. 714 00:29:40,007 --> 00:29:40,907 Mange, thanks for joining me. 715 00:29:40,907 --> 00:29:43,847 It was really great getting to know you and super cool to chat about the product. 716 00:29:44,087 --> 00:29:45,407 I'm excited to see where it goes from here. 717 00:29:45,872 --> 00:29:46,922 Thanks for having me on the show. 718 00:29:46,922 --> 00:29:47,607 It was a lot of fun. 719 00:29:47,887 --> 00:29:48,907 Of course, we'll catch you soon. 720 00:29:48,907 --> 00:29:49,192 Take care.