IMO that would clutter the results too much.
Gamification
Make PriEco fun to use through achievements and rewards.
This seems out of place for a search engine?
3 Likes
This new index doesn’t update so precisely and updates every ~4M inserts
I’d say that’s good enough especially as it may already have crossed 400M
Date is a very asked for feature
I’ve even got here a mockup where date was next to the result data. I am not sure if all they wanted was a filter or something more with it. I personally wouldn’t need it next to the result. Well maybe in this kind of way
Gamification idea is old but I meant something like this
or what Ecosia mainly does
The things there are just ideas, sometimes years old. I wouldn’t be implementing most of them at this moment and appreciate user feedback to know if some of them would be very liked (implement them soon) or very disliked (never implement them)
Regardless, Gamification is the most hated idea. The images got the second most dislikes which I find weird as I want PriEco to have images
Then the large ideas (Deep search, PriEco Premium, Interactive descriptions and Theme store) were more mixed
2 Likes
I personally would only want to filter the results for a specific range and would not really care when this specific result happened. It would make more sense to possibly filter from oldest to newest instead.
Very odd, since this is pretty much a must have 
3 Likes
Great ideas as most search engines just have a filter for the last year, month…
I can easily make it custom range too
Thank you for the great ideas! Will do both of them
Yes, here I’m thinking it maybe got dislikes for how it was framed?
PriEco did have images in the past
About them I have the best memory of a user who switched to PriEco from 4Get because of how images looked. They sent me this 4Get:
PriEco:
I’m certainly not in a position to start building own image index. So, they have to come from somewhere. I know that Resulti guy for 3 years, and he offered to help me get the images back to PriEco
2 Likes
I was confused why it said “4Chan” before the edit 
I agree with that user, since your image layout looks very clean
4 Likes
Notice the improvement in the languages, you’re doing a good job.
I noticed the photo on the information panel on the right, and it looks decent without breaking the visible search data.
The map you have built is useful; however, what I don’t like is the “AI, Sponsored products”, at a minimum. There is a problem here: the like and dislike buttons can be used by bots or illegitimately (abuse) to implement certain features. You’ll have to know how to differentiate between the legitimate and illegitimate ones; unless you choose to integrate measures.
I support donations, it’s better than “premium”.
Keep up the good work on development.
2 Likes
Thank you! Your English here is pretty “choppy” so I’ll try to answer based on how I understood your comment.
Thank you for complimenting the language, information panel and map!
If you’re talking about the roadmap:
It is too soon to be implementing them. PriEco is going to need more predictable source of income. For them I have in mind the Yadore company. They are from Germany. Have a very reasonable privacy policy. AND pay per click, so until you click on their product, they have no information. Not your IP, anything. By their privacy policy they collect truly only the minimum amount of information to have some certainty the paid clicks aren’t bots/spammed
I don’t mean LLMS as the ones Google and others are implementing. By this idea I’d like to let user specify by a longer text what they are searching for. It’s for when you don’t know name, or if the product/thing/location you’re searching for exists. But you can describe what you want to find and PriEco would look for product that matches the closest what you wanted. I came up with this idea when I needed something for PriEco but it was super specific and I didn’t know if it even existed but I could very well describe what I wanted. Normal search engines were awful for this. I found AI search engines: Google (AI mode), Ask Brave… better for it. Their issue was that they buried the results under a lot of text and I wasted a lot of time reading the slop text before I decided it sounded interesting, opened the source and had to fact check it again. I had to do this dozens of times.
I believe something that would well understand what I was looking for and search for it for me like AI search engines can do but show it in a classic list of results for a quick scan through them? I think it could be very useful
Right now I literary just get a Signal message a feature got like/dislike. I am well aware it has no bot/spam protection. But for now it doesn’t need it.
Could I ask you for one? 
Premium would be a way to support PriEco financially + it’s a different kind of web search engine. It has been done before but i’m up for a challenge and am willing to try to make it better than others. But will see, it’s just an old idea
3 Likes
You’re welcome. Thanks for letting me know about the language (English translation). Let’s take a look now, and hopefully this comes across better (improved prompt for the GLM 5.2 AI - max).
-
I understand the part about sponsored products. The whole clicks issue is a sensitive topic, but wouldn’t it be better to have a visible direct donation button? That would cover a lot of bases. Have you tested out that German company you mentioned?
-
Wouldn’t it be more viable to make it more direct and in-depth by default, even tailored to the date along with its context?
For example, I’m looking for the actual benefits of lemon on the skin based on facts—not generic claims or industry-manipulated science.
-> The results might be hard to find, but if it works the way it should based on context—staying true and delivering high-quality information on the topic when searched—that’s much better. I know exactly what you mean: cutting out the noise, the generic, the artificial, and all that stuff that isn’t what people are actually looking for.
-
Sounds good. I did that a while ago (the likes and dislikes).
-
No. I don’t use cryptocurrency systems (I don’t support them; I just stick to the basics).
-
Put it to the test. If you’d like, let me know so I can check it out, and I’ll put together a brief report to help you out.
One more thing: I’m on board with adding images.
1 Like
It is better but wouldn’t a normal translator be even better?
I have them in mind, but especially I wouldn’t be integrating advertisements without consulting it with the community
I do want to add it, I’m just too deep into Index and Ranking work, soon publishing an update in this forum. But certainly want to encourage donations
I am strongly improving the base ranking at this moment. I do want to have solid index first
PriEco did return pretty nice results. But you’re right, after I spent more time on the first one. I realized it was a company selling products. It surely is going to get better overtime. I’m mainly glad how relevant results PriEco is returning.
THIS RESULT CAME UP WITH USA LOCATION AND ENGLISH SET, your experience may vary
I’ll be adding a card payment as a donations option when I’ll be working on the donations page
PriEco Premium is for much later
Almost forgot about it, today I got a big part of the other person’s work. Maybe I’ll add them today
1 Like
If I use it normally, there’s a good chance of misinterpretations and words being added that I never said. I like to maintain as much control as possible to prevent that from happening as much as I can. In this message, I’ll use Deepl (its AI).
All in one (direct reply): What I mean is, keep moving forward and take your time. Keep me posted on updates so I can stay in the loop.
Keep it up!
3 Likes
I’ve added donations to PriEco
Supported are credit/debit card payments, PayPal, BTC, ETH, SOL and XMR
I’d really appreciate your donations
5 Likes
5 Likes
New updates are in PriEco!
Search suggestions
They come from DuckDuckGo and are proxied
-
You can use your mouse and click on search suggestion you like to make the search
-
You can use arrow keys to select different suggestions
- Once selected you can press tab to change your query to the suggestions, and it instantly shows new suggestions
- If you press enter it makes a search using the suggestion
-
As a creative additional, it adds to the search box the first suggestion and at any point while typing, you can press TAB to fill it in
Result information
Each result now has 3 dots next to it and by clicking them you get access to more information
- Action buttons
- Cache is how PriEco sees the result
- Internet archive button
- (I want to add there a button for getting a privacy report about the site using https://sitebehavior.org/ when the author becomes active online again)
- Speed, how hard the page is to read, classified intent of the page and from which index the page came from
- Text from the page
Design of this can be improved if you have ideas
Ranking
- New, much better semantic search
- Page usability metric
- Page relevancy metrics, the biggest one is the intent and matching user query intent to page intent
And many fixes and improvements
5 Likes
Bangs update!
No-js bangs now work
As well as PriEco exclusive feature called multibang
It is explained in this 2y old video https://www.youtube.com/watch?v=RRhzXh7e52E
5 Likes
I really wanted to bring them back. Now you can add shortcuts under the search bar. This is helpful if you have PriEco set as a new tab page
But these are small improvements. I just want to bring every single time some treat
Current situation
I must say I’m satisfied with what I’ve built. Yes, results aren’t “Google” level yet, but it’s improving.
Even I sometimes rely on Bangs, but I use them fractionally compared to the start of the index life
Security
In the last 5 days PriEco source code has been reviewed twice. Both times I wasn’t aware of it until the reviewer told me
1.time it was to ensure security of the codebase. I did get a report and some points were really serious. All extremely serious are patched now, and I’ll be moving on the rest.
The researcher was able to do a port scan, read a big chunk of my server file system, get running dockers with information about them, even .env file which exposed API keys, a phone number and other secrets. Luckily it wasn’t exploited, rather reported.
Even with this large access to my secrets, the reported had nothing to say about a difference of what is in PriEco Codeberg repository and what I’m running on the server. Rather they matched (at least in behavior) + I compile the project on the server and there is the source code sitting even at this moment. Not sure if they got access there, or bothered with matching it
2.time It was for an upcoming PriEco review
The reviewer told me there are too positive and too negative points, but the source code is getting overly positive score
Future
There is so much in making. I’m not sure if I should spoil it here. Maybee, it would be useful to get gradual comments. Not sure, tell me please
6 Likes
I love spoilers
2 Likes
Alright! Thanks for letting me know. I’m going to spoil it all for you
Main goal: Users
PriEco is pretty old already
Backstory
For the first 2.5 years I was relying on external indexes and hoped that with qualities: private, eco, open source + mixed results as back then SearX(NG), but now even Kagi and Uruky I would attract users. It actually wasn’t going bad and one of the biggest assets from this era was a big number of widgets from which many aren’t in current PriEco yet.
But I quickly found out I can’t rely on external APIs, not earn money and grow user base. It created a chicken and an egg problem:
- How to earn money with small user base
- How to grow user base when APIs get rate limited
- How to pay for API calls when I don’t earn
I decided to drop everything and create own index. On the crawler I’ve spent 1.5 YEARS!
add 2 months on index and 6 months on the rest and we are at the current moment.
I need to now figure out how to grow PriEco user base. It isn’t as simple as before as the results quality isn’t big tech level, so it isn’t an easy drop in replacement for most people, but at least now it’s scalable.
There are still many options I could use so hopefully it works out.
Don’t take me wrong, I like this community the most, but it’s either not huge or not that many of you use PriEco.
Main developmental goal: Decentralization
It all slowly progressed to decentralization, and now I see it as the natural next step.
Users: Here I want to upgrade PriEco Web Discovery extension to being similar to Brave Web Discovery Project + additions as mini crawler in the extension and query log scraper (As Brave + Mwmbl have)
All of it will be disabled when installed the extension and user enables what they are comfortable with.
Benefit of it is that it’s browser-agnostic. Unlike Brave’s implementation.
Brave search really laid out an outstanding theory on how they protect user privacy, but there is a hole. It’s a small hole and they even acknowledged its existence: “we still are able to find URLs that we rather not receive, we are talking about single digit figures in millions of records.” They are talking about URLs that are technically public, but search engine shouldn’t be able to find and index them. Examples are “unlisted” YouTube videos, collaborated documents (Google Docs…) and mainly shared files from cloud storages.
We had a long discussion and we haven’t figured out how to prevent this too for now. At the worst case the extension is going to have a warning about this. Not sure if Brave discloses this hole somewhere.
Nodes
There is no documentation for now. But PriEco already includes “mini crawler” So that anybody can run PriEco on their machine and support the index. (For now it isn’t that simple and if you’d like to, either wait and tell me about it please or tell me and I can set it up for you)
But when a node runner already runs PriEco for the mini crawler, why not expose a port, purchase a domain and run own instance of PriEco? (Even freely modify the design, just follow the license)
Here is the problem of index, right now it seems like it is going to be quite complicated in terms of architecture to ensure the best value and quality out of it.
But for node runner it will be important to know that they need to be beneficial to the network. By crawling and/or by paying for results, just like API customers.
Collaboration
Recently I’ve contacted 1 very talented developer who proposed decentralization to a different web search engine but got rejected. I found his ideas and other github projects great. Right now we are exactly deciding how the decentralization should exactly work, so that we can collaborate and prevent disagreements
His, in my view, most amazing point I haven’t even thought about was that if we get enough and stable nodes we can shard the index much more. The huge benefit is that PriEco wouldn’t be bottlenecked by my ability to satisfy its growing storage needs
Crawling
Now back to core centralized crawling. There is another user whose goal was to help another web search engine accept user contributions. But again he was rejected. Luckily, we have PriEco. His points are similar to what I’ve talked in decentralization, but he seems it almost completely from user POV.
Ranking
I’m joined in Mwmbl and Marginalia search communities. Sometimes I “anonymously” shared knowledge there. But, I found Mwmbl founder really nice and it was his idea + I’m there to follow through on the idea.
We want to together create a relatively standardized and hopefully unbiased results quality test for search engines. Test them all and create a leaderboard.
The idea isn’t very though through yet. Mwmbl already has a pretty nice test he shared with me. But there is still a significant bias towards how Google ranks results.
These are the main points
3 Likes
@asciimoo Maybe I can help you with some search engine tips.