跳至主要內容

Rank documents by relevance with AI

Create text embeddings and rerank documents by relevance to improve semantic search.

MCP / oo CLI + Skill2 個接入服務

可以做什麼

Connect a query with the most relevant passages in a document collection. Generate text embeddings for similarity workflows or rerank candidate documents around a research question.

Rank documents by relevance

Create text embeddings and rerank documents by relevance to improve semantic search.

Search team knowledge

Retrieve relevant passages from indexed team documents.

試試這些提示詞

替換方括號裡的內容,複製到已接入 OOMOL 的 AI 助手。

選擇連接方式

Rank documents by relevance

Create text embeddings and rerank documents by relevance to improve semantic search.

Use OOMOL MCP with AIsa to complete this task: Rank [documents] by relevance to [query] and show the best matching passages.

Search team knowledge

Retrieve relevant passages from indexed team documents.

Use OOMOL MCP with OOMOL RAG to complete this task: Use OOMOL RAG to find passages about [question] in the team knowledge base. Cite the returned sources.

開始使用 Semantic Search

  1. 接入你的 AI 助手

    選擇連接方式

    在 MCP 用戶端中新增 OOMOL,使用你的團隊完成登入,即可呼叫接入服務。

  2. 選擇服務

    AIsaUse OOMOL-provided access or connect your AIsa account.

    Create text embeddings and rerank documents by relevance to improve semantic search.

    查看 AIsa 接入說明 →

    OOMOL RAGSign in to OOMOL; no separate provider credential is required.

    Use OOMOL services with your current team.

    查看 OOMOL RAG 接入說明 →
  3. 嘗試第一個任務

    Use OOMOL MCP with AIsa to complete this task: Rank [documents] by relevance to [query] and show the best matching passages.

提供 Semantic Search 能力的服務

依任務選擇服務。操作數為各服務的全部操作總數。

所有操作

3 個工具

Create Jina Embeddings

Compute Jina embedding vectors for supplied text without storing a retrieval index. AISA documents 1024-dimensional output for the supported models; billing is token-based.

Usage-based pricing.

aisa.create_jina_embeddings

Rerank Jina Documents

Rank supplied documents against a query using a Jina reranker, preserving original indexes for retrieval pipelines. No provider index or stored document collection is created; billing is token-based.

Usage-based pricing.

aisa.rerank_jina_documents

Retrieve

Retrieve the most relevant text chunks from the current OOMOL team's knowledge base.

查詢向量化: 100 點數 / 1M Token; 重排(預設開啟): 100 點數 / 1M Token
查詢向量化
100 點數 / 1M Token
重排(預設開啟)
100 點數 / 1M Token
oomol_rag.retrieve

常見問題

How does document reranking help search?

Reranking orders candidate documents by their relevance to a query, helping surface the most useful passages.