Skip to content
SPECIFICTION
Get listed

Graphiti

github.com

Free plan APISelf-hosted
Visit github.com
Spec score 8.5 #7 of 26 in Retrieval-Augmented Generation Tools

What it is

Graphiti is an open-source framework for building and querying temporal context graphs for AI agents. It records validity periods for facts, allowing applications to ask what is true now or what was true at a previous point in time. New structured and unstructured information is integrated incrementally rather than requiring a complete graph rebuild. Derived entities and relationships remain linked to their source episodes for provenance. Retrieval combines semantic search, keyword search, and graph traversal, with result reranking. Developers can define entity and edge types with Pydantic models or let structure emerge from incoming data. Supported graph backends listed by the project are Neo4j, FalkorDB, and Amazon Neptune; Kuzu is deprecated. Graphiti includes FastAPI REST and MCP services, supports several LLM providers, runs locally, and is self-hosted. It requires Python 3.10 or newer and an LLM API key. The Apache-2.0 framework is free.

Spec sheet

What it does
Graphiti is an open-source framework for building and querying temporal context graphs for AI agents.Read 4 Oct 2026 · github.com
Purpose
Graphiti is an open-source framework for building and querying temporal context graphs for AI agents.Read 9 Oct 2026 · github.com
Deployment
The maker’s Graphiti page describes the framework as runnable locally and distinguishes it from Zep’s managed commercial platform.Read 4 Oct 2026 · getzep.com
API
The repository describes a REST API service built with FastAPI.Read 4 Oct 2026 · github.com
License
The GitHub repository identifies the project as Apache-2.0 licensed.Read 9 Oct 2026 · github.com
Changing facts
It tracks when facts become valid and when they are superseded, preserving historical information for point-in-time queries.Read 4 Oct 2026 · github.com
Custom ontology
Developers can define entity and edge types with Pydantic models or let structure emerge from data.Read 9 Oct 2026 · github.com
Data and ontology
It incrementally integrates structured and unstructured data and supports developer-defined entity and edge types using Pydantic models.Read 4 Oct 2026 · github.com
Database integrations
The README lists Neo4j, FalkorDB, and Amazon Neptune as supported graph backends and marks Kuzu deprecated.Read 9 Oct 2026 · github.com
Deployment and support
The README describes Graphiti as self-hosted and self-managed, with users responsible for building and operating the surrounding system.Read 9 Oct 2026 · github.com
Graph databases
The repository lists Neo4j, FalkorDB, and Amazon Neptune as supported graph backends, and marks Kuzu as deprecated.Read 4 Oct 2026 · github.com
Graph updates
New data is integrated incrementally without requiring complete graph recomputation.Read 9 Oct 2026 · github.com
LLM integrations
Graphiti defaults to OpenAI and also documents Anthropic, Google Gemini, Groq, Azure OpenAI, and OpenAI-compatible providers.Read 9 Oct 2026 · github.com
MCP
The repository includes an MCP server for episode and entity management, relationship handling, semantic and hybrid search, group management, and graph maintenance.Read 4 Oct 2026 · github.com
MCP and REST
The repository includes an MCP server for assistant integrations and a FastAPI REST service.Read 9 Oct 2026 · github.com
Provenance
Entities and relationships trace back to the raw episodes that produced them.Read 9 Oct 2026 · github.com
Requirements
The README lists Python 3.10 or higher and an LLM API key among its requirements.Read 9 Oct 2026 · github.com
Retrieval
Hybrid retrieval combines semantic embeddings, BM25 keyword search, and graph traversal.Read 9 Oct 2026 · github.com
Security and support
The maker’s comparison says Graphiti is self-managed and self-hosted, while enterprise SLAs, support, and security guarantees are features of Zep.Read 4 Oct 2026 · github.com
Security reporting
The security policy directs vulnerability reports through GitHub Private Vulnerability Reporting and lists version 0.x as supported for security updates.Read 9 Oct 2026 · github.com
Telemetry
Graphiti collects anonymous usage statistics, and its README says telemetry can be disabled with GRAPHITI_TELEMETRY_ENABLED=false.Read 9 Oct 2026 · github.com
Telemetry data
The README says telemetry does not collect API keys, graph content, queries, or personal identifiers.Read 9 Oct 2026 · github.com
Temporal facts
Facts have validity windows, so users can query what is true now or what was true at a point in the past.Read 9 Oct 2026 · github.com

Plans and prices

Graphiti (open-source) Free Apache-2.0 licensed framework · self-managed Read 9 Oct 2026 · github.com

Every Graphiti plan and price →

Strong points and limits

Strong points

  • Temporal validity enables point-in-time queries
  • Incremental integration avoids full graph recomputation
  • Hybrid retrieval with result reranking
  • FastAPI REST and MCP services included
  • Apache-2.0 open-source license

Limits

  • Requires Python 3.10 or higher
  • Requires an LLM API key
  • Self-managed operation is required
  • Kuzu backend is deprecated

Who it suits

Graphiti suits developers building AI agents that need temporal facts, provenance, hybrid retrieval, and custom graph schemas. It is intended for teams able to operate a self-managed, self-hosted framework.

Verdict

Graphiti combines temporal tracking, provenance, and hybrid retrieval in a free framework for AI-agent context graphs. Its self-hosted model and Python and LLM requirements make it a fit for teams prepared to run the surrounding system.

Where Specifiction places it

Alternatives to GraphitiAll alternatives →