Answer Socrates generates hundreds of questions for almost any topic, in many languages. You can use these questions to write authoritative, thorough content, to brainstorm product ideas, and see what real Google users are searching for.
@goncalo_pimentel@vladojsem maybe, but when I tried using Portuguese, it didn't work. The UI/UX are nearly identical. I don't get some people. If you're going to do this, why not try to innovate and create a better product?
@goncalo_pimentel@vladojsem It is quite similar, yes (as Ahrefs was to Semrush or Facebook to MySpace). Also check out Ubersuggest and the very cool
Scout Suggest. And although we do have some interesting, unique features already, like PAA, trends, performance, often more results, we don’t consider that differentiation as a necessary justification for more competition to exist.
So while yours is a perfectly valid criticism, it’s a common fallacy when looking at a new project or startup to see it as the end result, rather than the beginning.
@ytspar thank you. makes sense. there are many ways how you can differ in the future. and yes, sometimes products seem to be similar, but there is a difference at the end. for instance, answerthepublic has only 2 free searches per day. you are also showing trends and what people ask. what i am missing is possibility to download suggestions into the spreadsheet. good luck with your product!
💻👷 Tech stack
🎨 Frontend
The frontend is implemented using Svelte and Sapper, going against the grain of popular frontend development and using a compiler to produce plain JavaScript that runs without any runtime dependencies, resulting in a small and snappy bundle. Styling is done using Tailwind 2.0, which has been a pleasure to use.
⬇️ Backend
On the backend a NodeJS server is gathering the data by scraping and storing it in a cache. The asynchronous nature of JavaScript makes it easy to run all scraping in parallel.
🌐🌐🌐 Scraping / Proxies
We developed our own tooling to make most efficient use of proxies, preventing rate limits by carefully planning how often any individual proxy is used. This allows us to keep operating costs low so we can offer a free product.
Cheap scraping at scale requires lots of "internet street smarts" not taught in school. We're considering building a product that intelligently selects the right proxy so that developers don't have to spend time benchmarking various subnets and cloud providers. APIs like this already exist but are too expensive to use at scale.
🌐 Network
Cloudflare takes care of SSL and caching. Using a Page Rule we're able to cache not only images but also API responses on the edge, for free. This Page Rule "Cache Everything" isn't talked about often, but in my opinion it's one of Cloudflare's best offerings. The app runs on a single bare-metal server in a single location, but data is cached in over 200 cities for a snappy performance from anywhere in the world.
❓What should we build next?
Answer Socrates is far from being finished. What feature would you like to see?
- JSON, CSV export
- More word modifications to generate even more content ideas
- Historical data & alerts when new content ideas are discovered
- An API to integrate the data into your workflow
- APIs to build similar products (proxy rotation, Google suggestion, translation, etc)
I’d love to see this integrated into a Quora bot lol. Would be cool to see how many of the questions actually generate conversations which could lead to prelaunching quality content.
@refuseillusion well essentially you create a bot to ask/submit questions on Quora then you set it to notify you if a question meets a certain threshold (let’s say 10 answers). Then, you take those top questions and create long form content for them.
@refuseillusion@jakecreps I see. That bot could also be used to participate in the Quora Partner Program, which pays for asking "good" questions. I'm intrigued that there is a market for that - a person that asks questions all day long but might not even be interested in the answers per se.
The questions currently generated are pre-validated in the sense that there already is search traffic for them
We might add Quora as a data source, scraping questions that have already been asked. Theoretically I'd assume that by the time a question shows up in Google Suggestions, someone already asked it on Quora too, but I could be wrong about that
Gleek
Gleek
Answer Socrates
Gleek
Answer Socrates
Answer Socrates
Answer Socrates
Answer Socrates