@focusmcp/fts

Full-text search with TF-IDF ranking — index files, search with relevance scoring, suggest completions.

Measured economy

Static benchmark — no LLM, no variance, reproducible:

| Tool | Native equivalent | Brick output | Savings | |---|---:|---:|---:| | fts_search | 5,560 tokens | 370 tokens | -93.3% |

Conditions: payload-only measurement (iso-call). Multi-turn agent dynamics not modeled. See global benchmark for methodology and full per-tool table.

Tools

| Tool | Exposed as | Description | |------|-----------|-------------| | index | fts_index | Index a directory for full-text search | | search | fts_search | Search the index with TF-IDF ranked results | | rank | fts_rank | Rank a specific set of files by query relevance | | suggest | fts_suggest | Auto-complete a partial query using indexed terms |

How it works

The brick maintains an in-memory inverted index. Indexing a directory tokenizes each file (splitting on non-alphanumeric characters, filtering tokens shorter than 2 characters) and builds a term → postings map.

Scoring uses classic TF-IDF:

• TF = term count in document / total terms in document
• IDF = log(total documents / documents containing term)
• Score = sum of TF × IDF for each query term

The index is reset on each fts:index call.

Supported file types (default)

.ts, .js, .md, .json

Custom extensions can be passed via the glob parameter (e.g. *.ts,*.py).