# Component Bge Reranker

> Rerank candidate documents by cross-encoder relevance score with a local BGE reranker model (BAAI/bge-reranker-large). Use after first-stage retrieval when semantic precision matters more than latency.

- Skill: `docailab/component-bge-reranker` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add docailab/component-bge-reranker`
- Raw SKILL.md: https://api.skillmd.com/api/skills/docailab/component-bge-reranker/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Web & Frontend
- Author: DocAILab (https://skillmd.com/u/docailab)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/docailab/component-bge-reranker

---


# BGE Reranker Component

Provide the `reranker` capability for an Agentic RAG slot.

## Interface

Input `RerankRequest`:

- `query`: original retrieval query
- `documents`: JSON-compatible documents containing `id` and `text`
- `top_k`: maximum returned documents
- optional `model`, `batch_size`, `max_length`, `device`

Output `RerankResult`:

- `documents`: documents re-ranked by normalized cross-encoder `score`

## Execution

Run `scripts/component.py:run(inputs, context)`. The Component lazily loads the local BGE cross-encoder and scores `(query, passage)` pairs with a Sigmoid-normalized relevance score. It performs reranking only; do not use it for first-stage retrieval or generation.