insensitive accent

Could someome lead me in how to use the asciifolding filter to make accent insensitive? like a search for “nao” and “não”. Where and how should it configure it??
Im trying to figure it, but im starting with elasticsearch.


If you use Fess 13.3, the step is:

  1. Download fess.json at Backup page
  2. Add asciifoloding into standard_analyzer in fess.json
        "standard_analyzer": {
          "type": "custom",
          "char_filter": [
          "tokenizer": "unigram_tokenizer",
          "filter": [
  1. Upload fess.json at Backup page
  2. Start reindexing at Maintenance page

Perfect. Thanks.