glossdex is private semantic search over a folder of photos and documents that returns nothing when nothing matches. A local model writes a short gloss for every file — a list of independent phrases saying what is in it — and glossdex embeds the gloss and its phrases, then decides at query time how many results the evidence supports. Often that is one or two, and sometimes it is none: ask a photo library for a giraffe and you get nothing rather than ten unrelated pictures. It needs no training, no click logs and no relevance labels, and nothing leaves your machine.
Be the first to rate Glossdex. Your feedback helps other users decide.
Let a local model write an independent-phrase gloss for each file, then embed it.
Ask a question and get only as many results as the evidence supports, sometimes none at all.
Pass a fixed top-k if you would rather always see the nearest matches.
Similar tools in the same category — compare side-by-side
RemainderBot spends the Claude or Codex quota you were about to lose. Your subscription resets every...
gen-system reviews code changes with facts from a parser rather than guesses from a model. For each ...
anygrep searches your notes, PDFs and screenshots by meaning, from the terminal. It runs one model, ...
Guides and comparisons about code tools
Ten launches that keep the model, the index and the facts on your own machine, from search that admits no match to GPU training in CUDA.
Ten Show HN launches that answer the same question differently: where does what an agent learned live, who owns it, and who can read it back?
Ten recent launches for running agents where they can do less harm: virtual machines, air-gapped sandboxes, sequencers, and control planes you host yourself.
Ten launches aimed at the same bottleneck: agents re-read their own history, and every re-read is billed. Compression, code mode, prices and measurement.
Ten recent MCP servers that connect agents to data and systems they cannot invent: B2B records, financial metrics, flights, CAD, and keyword volumes.