graphiti
от emasoudy
Personal DevelopmentКратко
Knowledge graph operations via Graphiti API.
Описание
Что делает этот навык
Provides knowledge graph operations through Graphiti's REST API. Supports searching stored facts, adding new episodes or memories, and managing entities and relationships in a graph database backed by Neo4j and Qdrant.
Combines graph-based relationship traversal with vector search, enabling semantic queries that return structured facts rather than raw text chunks.
Когда его использовать
- Searching past project notes stored as graph facts
- Recording meeting summaries as new knowledge episodes
- Querying relationships between people and projects
- Persisting AI conversation memory across sessions
- Retrieving context about a specific topic from accumulated knowledge
Пример сценария
User asks: What do we know about the Phase 1 integration?
- 1
Resolve Graphiti service URL via config, env var, or localhost default
- 2
POST query to /facts/search with the relevant search terms
- 3
Receive ranked fact results from the knowledge graph
- 4
Return structured facts to the user
A list of relevant stored facts about Phase 1 integration, including relationships and entities extracted at ingestion time
Требования
- Neo4j database instance
- Qdrant vector search instance
- Graphiti service running and accessible
Установка
clawhub install emasoudy/graphitiВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.