Mneeemo search uses record titles, URLs, domains, descriptions, notes, and tags. Start with a short, distinctive clue, check the result, and try another word if necessary. It is text search: it does not automatically read the whole original website.
1. Choose a clue from a searchable field
One link can be found through different paths. The title identifies the source; the note preserves your intention; tags describe characteristics; the domain helps when you remember the site. The description depends on recorded preview text, which may be incomplete.
Words found only on the destination page do not automatically become searchable. If you want to find a reference through a specific idea, record that idea in a note or another appropriate field.
- Title: search for a word identifying the topic.
- Note: search for the task or decision you recorded.
- Tags: search for a recorded characteristic.
- Domain and URL: search for a distinctive part of the address.
2. Try one word before a long sentence
Search normalizes capitalization and accents and has limited tolerance for small typing errors. This can help with accented words, but does not make it semantic search or guarantee that every typo will be understood.
Terms in a query can match different fields. A query with several words may also return partial matches; do not treat the result as proof that the entire phrase is in the record. Open and check the reference.
3. Understand server search limits
The search endpoint, also used by the MCP integration, accepts queries of 2 to 200 characters. It examines up to 1,500 recent records outside the trash and returns up to 150 results. If the library exceeds that window, the response indicates that the search was truncated.
An older item missing from this search has not necessarily been deleted. Explore its folder, the library listing, or other clues. Do not use search counts as a complete inventory of a library larger than the scanned window.
4. Investigate an unsuccessful search
Check the account, active filters, and trash. Try the domain or a more distinctive title fragment. If the detail you remember was never recorded, look through the project folder instead of repeating the same sentence.
Once you find the link, add the missing clue and repeat a short query. Improve a real path back to the source rather than accumulating words that make the note hard to read.
A useful query uses recorded clues; a useful note preserves the clues you plan to use later.
Questions you might have
Does search read text inside images or videos?
Not automatically. It uses the record's text fields, without assuming OCR, video transcription, or full-page indexing.
Can I search without accents?
Yes. Normalization removes differences in accents and capitalization. The clue still needs to match searchable fields.
Why does a result not contain every query word?
Ranking allows partial matches according to the number of terms. Try a more specific word and check the records before interpreting the result.
Put the idea to work in your library.
Start my library