MeiliSearch: Ultra relevant and instant full-text search API

github.com

4 points·NetOpWibby··2 comments

After installing, I was able to access the API from the localhost. But, without any firewall active on the host, I wasn't able to access the API from other machines in my local network. The documentation didn't mention how to enable full network access to the API.

Also, it would be a good idea to explain the scope of the project in the documentation. Apparently this isn't a tool that indexes a bunch of PDF, DOC or TXT documents on a local computer, or HTML documents on a web server. It expects some form of pre-indexing, like the movie list in JSON format in the example. The "full text" aspect is relevant to that list or pre-index, not the full text of the referenced files, right?

Hey, I am in the MeiliSearch team.

You can enable your full network access to the API by launching it like this `meilisearch --http-addr=0.0.0.0:7700`. The default listening address is 127.0.0.1 which is why you can only touch it from localhost. You can also pass it as env variable, you can read more on the documentation https://docs.meilisearch.com/advanced_guides/binary.html#env...

And yes, MeiliSearch does not ingest PDF or DOC out-of-the-box, it's a RESTful API to which you can send JSON. You might want to look at https://github.com/dadoonet/fscrawler, but it is working only with Elastic atm.

Reply on news.ycombinator.com