Vector Memory System Integration

knowledge skill

Indexes and retrieves text chunks using vector embeddings. Produces a searchable index of documents with similarity scores for queries.

Worked examples

• {"input": {"documents": ["The quick brown fox jumps over the lazy dog.", "Lorem ipsum dolor sit amet."], "embedding_model": "text-embedding-3-small", "chunk_size": 100, "chunk_overlap": 20}, "output": {"index_id": "idx_7f3a9c", "index_size": 2, "embedding_model": "text-embedding-3-small"}}
• {"input": {"documents": ["Agent skills are defined operationally.", "JSON schemas must be concrete."], "embedding_model": "all-MiniLM-L6-v2", "chunk_size": 50, "chunk_overlap": 10}, "output": {"index_id": "idx_b4e2d8", "index_size": 2, "embedding_model": "all-MiniLM-L6-v2"}}

Input

• documents: list[str]
• embedding_model: str
• chunk_size: int
• chunk_overlap: int

Output

• index_id: str
• index_size: int
• embedding_model: str

Details

• Skill type: knowledge skill
• Safety level: safe_public_research
• Version: 1.0.0