What is Teclis?
Teclis is a search engine for discovering attention-grabbing, routine results on
‘neat’ net sites.
Teclis is no longer a Google replacement, and works simplest
for study and discovery with
wide(er) search phrases love the examples above. For a stout Google replacement,
that includes Teclis, take a look at Kagi.
Form a requirement to look. Exhaust the examples above for a few suggestions.
To allow experimental semantic search mode (search in natural
language) plot the demand in ( ), for instance (I want to look at search
OpenSearch specification is supported for more uncomplicated addition as a
search engine in love minded browsers.
Gain a requirement that has only one lead to Teclis! Then read
Discovering estimable yelp on the online grew to changed into extremely laborious, as if the
ad-monetized search engines like google and yahoo
are tuned to handbook the consumer in opposition to the net sites that
feed into the identical ad meals chain. As a outcome, the identical “high” net sites dominate the hunt results, fortunately patting every different
on the support.
The remainder of the online has extra or less long gone incognito.
Teclis is an
are trying to ground the less known net, the online of creativity and
self expression, the extra humane net.
crawler is hybrid, using async python requests and puppeteer with uBlock Starting assign aside. The technique
detection works is we rely the
settle on of uBO blocked requests on the page, and if too many (threshold is made Up our minds
to 5), we kick it out, leaving only “neat” pages
in the index.
is additionally routine in a single design that this would possibly occasionally maybe well be conscious an provocative uninteresting hyperlink to its net
archive page, making an are trying its simplest to withhold the page in our
index (you will gape those results below “Net Archive” piece).
To invent search results, triple rankings with stout-text search (through Elasticsearch / Typesense) and NLP-based mostly
semantic search (using sentence
transformers to invent embeddings and SCANN
to look through vector home) are mixed to invent final
ranking of results. Responses are served using FastAPI.
Teclis additionally makes utilize of results with permission from Marginalia
Search (one more noncommercial search engine).