Project: Preturnio - a full text search engine for Bitcoin

[preturnio]

Hi. I wanted to share my new project Preturnio - a full text search engine for Bitcoin. With Preturnio you can search for text inside every OP_RETURN, Coinbase transaction and ordinal transaction since the genesis block. You can play with it over at https://preturnio.com

Block explorers can search for transactions, blocks and addresses but searching for arbitrary text is a challenge. Data is stored in the blockchain natively in hexadecimal format and not easily readable by humans. Preturnio solves this with an easy-to-use search interface. Simply enter the text you’re looking for and Preturnio will return highlighted results found in any OP_RETURN, Coinbase or Ordinal transaction. See when the data was created, which block it was confirmed in and other relevant details.

You can search for anything: source code, names, quotes, brc-20 tokens – any plain text you can think of, you can find in the blockchain with Preturnio. You can also share transactions with a direct link or even a preview image. Posting to Twitter will automatically create a Twitter card.

Ordinals are of course supported. Each inscription can be previewed inline or full screen. The platform support all transactions encoded with text-supported MIME types and we categorize inscriptions into two buckets – text and HTML. You can toggle to source code view with a familiar search interface to find the text you’re looking for inside the inscription payload. When you share an inscription, Preturnio will try to include a preview of the inscription whether it’s text or HTML. Once you’ve found your inscription, you can then search inside it with a familiar ‘Find’ experience to look for specific text.

Naturally each search result can be verified with a single click to on one of several public block explorers as well as ordinal explorers for inscriptions. It's also easy to copy the TX ID to verify on your own node.

Preturnio has also cross referenced all transactions against a published list of all known English words (more languages coming soon) so you can browse individual words stored somewhere on the blockchain. This is actually a lot of fun!

Check out the https://preturnio.com/docs to learn about other search features supported.

Feedback is encouraged and appreciated. Thanks for listening!

[Yamane_Keto]

I liked the idea, the design of the site is elegant and clean, the speed of response and the way the data is displayed, it is more of a professional work than a side project.

I will bookmark this page.
The only thing I didn't like is that my search results and others appear in the bar below.
I suggest adding random words in the search option so that it helps the user if he has no ideas.

What are your future plans?

[hugeblack]

The project looks complete, the design is cool and the search is smooth, why not create an ANN Topic here ----> https://bitcointalk.org/index.php?board=84.0.

Suggestions, um let me think about the following:
 - Dark Mode.
 - OP_RETURN used in Bitcoin history^[1].
 - Search for information by an individual such as Satoshi or a mining pool.
 - Proof of Burn addresses, such as messages sent through addresses.

[1] https://twitter.com/pourteaux/status/1361821176801157122

[preturnio]

I liked the idea, the design of the site is elegant and clean, the speed of response and the way the data is displayed, it is more of a professional work than a side project.

I will bookmark this page.
The only thing I didn't like is that my search results and others appear in the bar below.
I suggest adding random words in the search option so that it helps the user if he has no ideas.

What are your future plans?

Thanks for the feedback.

This is a side project - I'm just super OCD about UI/UX and spent a lot of time trying to make it smooth and simple to use. I guess that's positive if people think it looks more professional

Regarding search results: that's good feedback. The label above the marquee says 'recent searches' when this should really be 'what are people searching for?'. It sounds like the concern here is one of privacy which is valid. I did think about a dropdown in the search text field similar to what Google does with 'recent' / 'trending' searches but I didn't think that made sense until I have a lot of search query history to work with, but I will think about how to improve this.

Plans are to optimize speed and improve indexes - for example, consider the following OP_RETURN:

https://preturnio.com/8bbc96506afcee124600c94dcdc85c619db96f825a2d065bccd0b3129962ed3c

If you wanted to find this transaction, you could search for the entire string or just '2565C53636AB9239157C9F7D2E97656866393A004DE00BAB742FD4F57DC66CB7' - but let's say you wanted to look for the last few characters (e.g. 4F57DC66CB7) - you won't get any results. So it would be good to add some options for more powerful searching. I also wonder if some people would find value in searching for the raw hex instead of the encoded data. Obviously the result set could be much bigger but would that be of value?

I would also like to support adding search to the mempool - just trying to figure out how to do that with reliable infra. The initial index was built using my home node but now I use Quicknode for each new block. Handling the mempool would, I think, require my own node and that's actually pretty expensive to host in the cloud so trying to figure that out. I am of course open to other requests from folks who have ideas to help improve the concept.

Thank you for the feedback and for spending the time to take a look at the project. I really appreciate that.

[preturnio]

The project looks complete, the design is cool and the search is smooth, why not create an ANN Topic here ----> https://bitcointalk.org/index.php?board=84.0.

Suggestions, um let me think about the following:
 - Dark Mode.
 - OP_RETURN used in Bitcoin history^[1].
 - Search for information by an individual such as Satoshi or a mining pool.
 - Proof of Burn addresses, such as messages sent through addresses.

[1] https://twitter.com/pourteaux/status/1361821176801157122

This is great - thank you! But so many questions...

Dark Mode - yeah. I'm really bad at CSS and while building the UI I found it very frustrating to continually tweak two different CSS fields but I really need to do this.

Regarding the other three suggestions, can you provide some more details?

1. OP_RETURN used in Bitcoin history^[1]: can you be more specific here?

2. So you can search for details related to a mining pool because I'm indexing coinbase transactions but the 'problem' here is performance. So let's say you wanted to find transactions with the word 'USA Pool' - you can search for that but Preturnio will find over 20,000 records and these searches are typically very slow. In fact, the word 'mined' is in my stop word list (try searching for the word 'mined' and you'll see what I mean).

3. Not sure what you're referring to here - can you give some examples? Remember that Preturnio is about searching for plain text - so if you're looking for specific 'types' of transactions that don't include plain text, I'm not indexing those transactions. I hope that makes sense and answer the question but an example would be helpful....

[examplens]

Dark Mode - yeah. I'm really bad at CSS and while building the UI I found it very frustrating to continually tweak two different CSS fields but I really need to do this.

I understand your frustration with CSS, I've struggled with it myself.
If you have a lot of work to switch between dark/night mode, maybe you can find some design in the middle, which will be satisfactory for both parties. The background does not necessarily have to be white, even if it was a light-mode version.

[paid2]

Very very cool! Your site is interesting and really well done. The website is fast, fluid and super responsive.
I agree 100% with hugeblack's idea for darkmode, even if from a personal point of view I use a plugin that transforms all my web pages in darkmode version.

Thanks for your work and for sharing it here. I will spend some time looking at the reference word list and try to do some research.
You can find everything it's really funny, some more "educational" than others:
https://preturnio.com/ordinal?txid=d8ef417a8575e0fa3da7ab11f9e91ef7582c7275f54f8b9dc587c98c676bd41c

Sorry OP, I don't have any smerits at the moment but I've added your topic to my list of topics that I'll merit later 

[ABCbits]

This is really cool project. I search few random things and found many things such as personal message[1], copy of Mike Hearn
blog[2] and even support to Ukraine[3]. Your project could be popular among Bitcoiner, NFT/Ordinal user and few people who looking for specific content.

[1] https://preturnio.com/8e801876a1c0aadbf4de88c3509aef0b1b61b104958888a54991c34c885eaedc
[2] https://preturnio.com/e0071e2c7ab66227feb4ccec8c9941edd07a01c6f05444bc5980fde7c955b860
[3] https://preturnio.com/073ffed8259b53a44bf8813e2d83c16d95ef792e08afb89855d19d367659cb34

Feedback is encouraged and appreciated. Thanks for listening!

• Dark mode and optionally dark mode by default. I know it's already mentioned, but IMO it's important feature.
• Better search algorithm which accept typo from. For example "block size" show 7 results while "block sizze" show 0 result.
• Warning about potential mature/controversy things on search result. For reference, there are 687 results when i use keyword "sex".

--snip--

3. Not sure what you're referring to here - can you give some examples? Remember that Preturnio is about searching for plain text - so if you're looking for specific 'types' of transactions that don't include plain text, I'm not indexing those transactions. I hope that makes sense and answer the question but an example would be helpful....

I believe he refer to old technique which use 20-byte hashed public key (on Bitcoin legacy address) as a way to store text and other arbitary data. Visit https://www.righto.com/2014/02/ascii-bernanke-wikileaks-photographs.html and check section "How to put your own message in the blockchain".

[preturnio]

• Dark mode and optionally dark mode by default. I know it's already mentioned, but IMO it's important feature.
• Better search algorithm which accept typo from. For example "block size" show 7 results while "block sizze" show 0 result.
• Warning about potential mature/controversy things on search result. For reference, there are 687 results when i use keyword "sex".

Thanks for these suggestions!

• Dark mode: working on it.
• This is a limitation of the search index engine and how it handles individual tokens. There are multiple 'strategies' to use and the one that fits most use cases looks like 'size' and 'sizze' as two different words. There's not much I can do about that without significantly changing the index configuration. I believe the solution might be to up the 'fuzzy' settings and I'll take a closer look at that - I do need to do better in terms of getting closer to a UX that folks expect from Google.
• That's a tough one. I actually agree with this concept but I worry that any type of messaging opining about content can put a lot of Bitcoiners off. Transparently: I am removing a bunch of 'inappropriate' words from the search marquee at the bottom (you won't see this if you're using Safari due to marquee performance issues on that platform) but if someone types in a word and that word is somewhere in the index, it will come back with results. Luckily I'm only dealing with text and not images. Additional transparency: I have a censor flag I can flip on a tx in case there is something that is clearly 'very very bad' - if I flick that switch for a tx, the encoding results won't be shown in the UI - so you'll still get a result - you just won't see the decoded text.

--snip--
I believe he refer to old technique which use 20-byte hashed public key (on Bitcoin legacy address) as a way to store text and other arbitary data. Visit https://www.righto.com/2014/02/ascii-bernanke-wikileaks-photographs.html and check section "How to put your own message in the blockchain".

Thanks for the reference - I took a look at that and I think I'll skip that for the moment - not sure the juice is worth the squeeze unless lots of people start asking for it.

[nioctib_100]

Dark Mode - yeah. I'm really bad at CSS and while building the UI I found it very frustrating to continually tweak two different CSS fields but I really need to do this.

I understand your frustration with CSS, I've struggled with it myself.
If you have a lot of work to switch between dark/night mode, maybe you can find some design in the middle, which will be satisfactory for both parties. The background does not necessarily have to be white, even if it was a light-mode version.

I've been doing web design for roughly the past decade, and I have say, I think your web design here is overall plenty fine. It's clean and modern, and in informational and data-heavy sites like this, you want to stay on the side of minimalism in design. If the design was flashy and animated, it'd confuse more "rookie" Bitcoiners to who a lot of your site's data is pretty foreign. Great site as all others have said, just wanted to give my two cents on the web design discussion.

[FatFork]

After using Preturnio, I have to say it's an amazing tool. It makes searching for specific text on the blockchain a breeze, and I love how it shows when the data was created and which block it was confirmed in. The word list is also pretty cool. However, it would be great if you could add a filter and sorting feature, so users can organize words by specific criteria.

Overall, great job OP! I really appreciate the effort you put into developing this tool.

[preturnio]

After using Preturnio, I have to say it's an amazing tool. It makes searching for specific text on the blockchain a breeze, and I love how it shows when the data was created and which block it was confirmed in. The word list is also pretty cool. However, it would be great if you could add a filter and sorting feature, so users can organize words by specific criteria.

Overall, great job OP! I really appreciate the effort you put into developing this tool.

Thank you thank you! Really appreciate the feedback.

Yes the platform needs better filtering and sorting for sure - I know a lot of people already want to filter out the BRC-20 token entries. Just need to figure out a nice UX for this.

Do you have any specific suggestions around filtering/sorting?

[preturnio]

I've been doing web design for roughly the past decade, and I have say, I think your web design here is overall plenty fine. It's clean and modern, and in informational and data-heavy sites like this, you want to stay on the side of minimalism in design. If the design was flashy and animated, it'd confuse more "rookie" Bitcoiners to who a lot of your site's data is pretty foreign. Great site as all others have said, just wanted to give my two cents on the web design discussion.

Appreciate that - not a designer myself - got someone to do the wireframes. Definitely tried to keep it minimal but also give it a little shine. I do think I need to fix the mobile UX - right now it's responsive but ends up just shrinking everything which isn't great - need to give this a little more thought to give the text some more prominence - this would obviously increase each result's real estate in the vertical but everything feels very squished on a phone now.

I appreciated your comment about using animation - I was a little conflicted at first about the animated logo GIF at the top and the scrolling marquee at the bottom - I know some people thing that's just dumb but I felt the animation was subtle enough (plus it's only on the home page). I kinda think the marquee is fun

[FatFork]

Thank you thank you! Really appreciate the feedback.

Yes the platform needs better filtering and sorting for sure - I know a lot of people already want to filter out the BRC-20 token entries. Just need to figure out a nice UX for this.

Do you have any specific suggestions around filtering/sorting?

You're welcome! I'm glad to hear that my feedback was helpful.

When it comes to filtering and sorting options for your platform, I would suggest taking inspiration from ord.io and implementing similar features. Ord.io has a clean and simple interface with filtering and sorting options that are easy to use. You can also consider adding additional filtering based on user feedback and preferences.

[preturnio]

Dark mode and optionally dark mode by default. I know it's already mentioned, but IMO it's important feature.

Dark mode is now live on the site. Thanks for the suggestion!

[dkbit98]

Good work, I think this tool can be useful.
I did quick search for several terms and some of them were Satoshi with lot of results, Theymos has a few spicy messages left, and phrase Bitcointalk gave me total result of 22, with 8 OP_RETURN and 14 inscriptions:
https://preturnio.com/results?p=1&b=bitcointalk

It would be cool if you could add alternative .onion domain, for people who are using Tor browser.

[preturnio]

It would be cool if you could add alternative .onion domain, for people who are using Tor browser.

Thanks for the feedback. I have no idea how to do this  but I will start researching how to offer the platform through a .onion domain.. great suggestion.

[so98nn]

Though I am not sure what real purpose could be (technically) but I’m sure it’s gonna be fun tool to be used for searching various layman topics and terms. When I opened your tool the first thought came in to search for ChatGPT and see what result it throws back.

I was expecting hundreds of thousand results in return but to my surprise it just came back with 22 results.

One noticing thing was, it associated the ChatGPT with one of the Darwins Evolution theory.

Question: does it hide text with asterisk for the safety reasons or for example out of copyright law to protect the breach of data?

In the result viewer shows data hidden with asterisks. (Many things to learn about this tool tbh)

# The State of Darwin’s Theory of Evolution in 2023

Gregory F. Coppola

**************Apocalypse Like Right Now**************

April 12, 2023

# Introduction

This essay explores the case *******against******* the (Darwinian) theory of ****************************************undirected evolution****************************************. We refer to proponents of undirected evolution as ********atheists********.

The conclusions are as follows:

- the default position should be ***********agnosticism***********, not *******atheism*******, until the data suggests otherwise, following the principle of **************maximum entropy**************
- the theories of ********************undirected evolution******************** (atheism) and ******************intelligent design****************** are exhaustive
    - i.e., there are only the two alternatives
- the theory of evolution is not like other theories in science
    - there are major “mysteries” remaining at every level of the atheist theory
        - the onus is on the atheists to fill out their theory
    - the theory of evolution is a research program not a theory
        - a research program cannot be falsified
        - the onus is on atheists to make falsifiable predictions, or admit that they don’t have a theory
    - evolution is mostly a set of “just so” stories, that superficially use the word “because” to describe the past, but don’t actually predict when evolution will take place
- we identify a fallacy, the ******************functional fallacy******************
    - the belief that a deep understanding of biological processes as they function today implies an understanding of the *******genesis******* of these processes
        - the everyday operation of biological functions **is** well-understood, but the genesis of these is not

# Only Two Options

We begin by discussing that there are apparently only two alternatives for the explanation of the human race.

We believe that this can be shown to be an exhaustive list.

The two alternatives are:

1. **intelligent design**
    - we are a product of design by an *****************intelligent being*****************
2. **undirected evolution**
    - we are a product of a series of **********undirected********** events
    - proponents of this theory are called ********atheists********

We believe that the definition of **********undirected********** that fits the intuition and practical usage of this term is:

- definition of **undirected**
    - arising **without** intelligence

Thus, as a matter of definition, there are the only two options:

- either we are a product of “an intelligent designer” or we are not

More precisely, consider any object $X$, which could be either:

- the universe
- the planet earth in its place in the universe
- any given species on planet earth
- humans

We say that either $X$ is a product of an “intelligent cause”, or else it is not. If $X$ is not the product of an “intelligent cause”, then it is a product of an “undirected” cause.

# A Priori, the Default Position Should Be Agnosticism

## Artificial Intelligence is a Program to do Science

There are certain areas of unclarity in the question of how science can be done.

We believe that in 2023 the philosophy of science should be informed by reference to artificial intelligence.

That is, artificial intelligence tools do *****science*****.

However, unlike humans, in order for a computer program to do science, it has to follow a specified procedure.

Thus, artificial intelligence programs force us to specify science as a procedure.

When humans do science, because the “software” for doing science exists in the human’s brain, a human being doing science does not need to *******consciously******* be able to explain optimal science.

Humans have been doing science since the first appearance of language.

But, the ability to describe science as an algorithm dates only the 21st century.

Thus, it has long been possible with **do** science, without exactly knowing what that actually means.

However, when a computer does science, because it is following a specific program, either the equations implemented in the program work, in which case artificial intelligence succeeds wildly, or they don’t, in which case the program can predict either nothing, or else nothing impressive.

Practice has shown that neural networks are trained by the maximization of certain equations.

## Maximum Entropy Training

ChatGPT is trained using cross-entropy using the principle of maximum entropy.

The principle of ***************maximum entropy*************** is:

- principle of ******************************maximum entropy******************************
    - find the most uniform distribution consistent with available information

In other words, given a data set, the principle of maximum entropy dictates we should pick the probability distribution that is the ************least biased************ ****************between predicting alternatives that the dataset does not support.

## Agnosticism

In the context of the debate between intelligent design or undirected evolution, ***if there were*** no evidence to choose between alternatives, the default position should be one of agnosticism.

However, in 2023, we do have abundant data-driven reasons to doubt whether undirected evolution is possible.

# The Necessity of Falsifiability in Science

## The Burden of Proof in Science

The way that science is ********supposed******** to work is that, in order to introduce a significant new theory, the proponent of that theory should make:

- clear, falsifiable predictions
- predict more than it mis-predicts

## Research Program versus Falsifiable Theory

However, there are certain cases in the history of science, where something else is able to arise: a ******research program******.

A research program is the *project* to look for a falsifiable theory.

A research program cannot be falsified.

## Chomsky’s Minimalist Program

Chomsky wrote of his “Minimalist Program”:

> It is important to recognize that the Minimalist Program (MP) under development in this work, and since, is a program, not a theory, a fact that has often been misunderstood. (The Minimalist Program, 2015, page vii)
>

In other words, *******nothing******* can falsify the minimalist program, because it is not a specific theory.

As the old saying goes:

> It’s good work if you can get it.
>

The idea that a scientist can be at work and taken seriously without a falsifiable theory is a deeply corrosive and
[/qoute]

[preturnio]

Hi OP, with recent BRC-20 and ORC-20 (both use/depends on Ordinals) hype, i wonder how your server/website process lots of small Ordinals transaction? I'm curious about it, so i decide to search few keyword related to those and it seems you haven't update your word stop. For example,
1. "orc-20" keyword return "32,355 results in 5.00 seconds".
2. "deploy" keyword return "30,181 results in 4.57 seconds".
3. "send" keyword return "699 results in 0.59 seconds".
4. "ordi" shows message "Your query took a little too long to process - to manage our limited resources, queries that take longer than approximately 5 seconds will time out. Please try a different query.".

Thanks for going into detail here. The explosion of ordinals has definitely been a bit of a challenge for this project - especially these BRC/ORC-20 tokens which have been inscribed so many times (as you probably know the provenance=timestamps are the key). So good catch as I had not run the stop-word analysis script and those terms should definitely be in there and have now been added.

When the project first started, BRC-20 wasn't a thing so looking in a block with ordinals you were going to find a nice collection of interesting stuff (text/html) - now (IMO unfortunately) if you're not doing a search for something specific and you just want to see what's in a particular block, you're not going to see those interesting text/HMTL inscriptions because you're going to be bombarded by BRC/ORD inscriptions. I should probably look at a filter so that you can find/browse the interesting stuff.

Thank you again for checking out the site and providing valuable feedback!

[preturnio]

One noticing thing was, it associated the ChatGPT with one of the Darwins Evolution theory.

Question: does it hide text with asterisk for the safety reasons or for example out of copyright law to protect the breach of data?

In the result viewer shows data hidden with asterisks. (Many things to learn about this tool tbh)

I appreciate you exploring the site! To clarify your comment - Preturnio didn't 'associate' anything - there is an inscription that was made in a specific transaction (e.g. 754f52551d5168972874cc9249ca0a5a4741cc610a3f48cdf0551e445a4ad1cc) where someone embedded that entire block of text inside a witness script (i.e. an ordinal) - that is the literal text that was inscribed. Preturnio is doing a text-search - i.e. pattern match - and it will return all Bitcoin transactions where it finds a match/partial match (see the Docs page for more details).

Preturnio doesn't mask any data - if you're seeing asterisks in the text, then they are literally in there. This is why I added the 'Verify' section of each result so you can go look at other Ordinal explorers to see the same result. You can also click on 'Inspect Inscription' if you wanted to look at the raw hex of the transaction.