FalkorDB
Open source graph database for AI agents, GraphRAG and multi-tenant workloads.
- Tool type
- Open source graph database
- Made for
- AI engineers and data teams
- Runs on
- Cloud service or self-hosted Docker
- Managed FalkorDB Cloud at app.falkordb.cloud, or run locally with Docker; API access is documented.
What it does
FalkorDB is an open source graph database that stores connected data such as tables, documents, code and events as nodes and relationships, queried with OpenCypher. Query results return connected facts and paths that can ground LLM answers, give agents relationship context, or support fraud, cybersecurity and supply-chain analysis.
Typical uses include GraphRAG retrieval, agent memory that persists relationships across sessions, and multi-tenant setups where thousands of isolated graphs run inside one deployment. The vendor lists integrations with LangChain, LlamaIndex, Graphiti and Mem0, and publishes its code on GitHub.
It suits AI engineers and data teams comfortable with graph queries or in need of relationship-aware retrieval. Access is through the managed FalkorDB Cloud, with a free-to-start sign-up, or by running it locally with Docker. Paid plans are priced per GB per month, and the performance comparisons against Neo4j are the vendor's own benchmark claims.
What it can do
- Stores connected data as a property graph queryable with OpenCypher
- Grounds GraphRAG retrieval in graph context rather than similarity search alone
- Provides long-term agent memory that tracks relationships across sessions
- Runs thousands of isolated multi-tenant graphs within one deployment
- Supports fraud, cybersecurity and supply-chain path queries across many hops
- Offers a documented API and open source code on GitHub
Works with
- LangChain
- LlamaIndex
- Graphiti
- Mem0
- Cognee
- g.v() visualization
Before you choose
- Paid plans (STARTUP from $73 per 1GB/month, PRO from $350 per 8GB/month) are published on the pricing page; ENTERPRISE is custom-priced via sales.
- Multi-tenancy and open source are stated across plans; TLS, clustering and high availability are stated as paid-plan features.
- Performance figures such as 496x lower P99 latency versus Neo4j are vendor benchmark claims.
Listed under
Get started
Start free on the managed cloud, or follow the local Docker quickstart.
- Sign up for FalkorDB Cloud at app.falkordb.cloud/signup, or follow the Docker quickstart at docs.falkordb.com/getting-started.
- Load connected data into the graph and query it with Cypher.
- Connect your AI framework (such as LangChain or LlamaIndex) or the GraphRAG-SDK, using the docs at docs.falkordb.com.
Plans and prices
Prices in USDSTARTUP is listed from $73/1GB/Month and PRO from $350/8GB/Month; ENTERPRISE is tailored pricing via sales. Free plan duration is not specified.
- Multi-graph and multi-tenancy
- Community support
- Multi-graph and multi-tenancy
- Graph access control
- TLS
- Automated backups every 12 hours
- Community support
- Cluster deployment
- High availability
- Multi-zone deployment
- Business hours support
Source: falkordb.com/plans, checked 29 Oct 2026. Prices can change; check before you buy.
Sources for this overview
Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.
- FalkorDB: Open Source Graph Database for AI Agents & GraphRAG ↗www.falkordb.com
- Company Information: Learn About FalkorDB ↗www.falkordb.com
- FalkorDB Pricing: Free Tier to Enterprise Graph Database ↗www.falkordb.com
- FalkorDB: Open Source Graph Database for AI Agents & GraphRAG ↗www.falkordb.com
Ratings and reviews
No ratings for FalkorDB yet.
What’s changed
Free accountSignals
Free accountData and security
Free accountCompare with similar tools
Free accountWhere this comes from
How we research- Company website
- falkordb.com: product, pricing and policy pages
- YC directory mirror
- Batch and company name
- Social accounts
- Accounts the company links, and its posts on X
- Human review
- Checked by a person
- knowledge_base
- manual_review
Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.