FalkorDB

Your AI coding assistant now knows FalkorDB. Say good-bye to: Wrong Cypher syntax, missing index types or unsupported clauses. These errors cost you debugging hours that never needed to exist - as well as costly tokens! Get the skills here: github.com/FalkorDB/skills

NEXT WEEK: The AI-Ready Data Summit - a free, live virtual event on March 31, 2026. 80% of AI projects never make it to production. Data issues are the #1 reason why. ai-ready-data-summit.com @lakeFS @g_korland

Give your OpenClaw agents accurate, structured memory By utilizing FalkorDB as the memory layer, you get: • Graph-based context (not flat vectors) • Per-user knowledge without complex filters • Better retrieval → more precise responses Check: github.com/FalkorDB/openc…

Re-watch FalkorDB cybersecurity webinar: 1) Ransomware spread sim via Monte‑Carlo UDF (host 12 90% infection). 2) Log4Shell dep graph with cycle‑safe traversal. 3) Max‑flow leakage model (10→52 Mbps). See UDF vs built‑in algo discussion. youtu.be/aFkgUT0ubYY

Join us for a technical session on advanced graph algorithms, with a cyber-security use case, user-defined-functions, and the recently-launched FLEX library by FalkorDB's CTO and Solution Architecture team. GMT: luma.com/m71fs7s7 PST: luma.com/1ewbzub1

NEW: FalkorDB now runs as a @Snowflake Native App 🏂❄️ You bind your tables, load data into a graph, and write Cypher queries: all inside your Snowflake account. No data movement or external infrastructure -> app.snowflake.com/marketplace/li…

Finally! A Text-to-SQL solution that actually works! (and it's open-source) When Text-to-SQL doesn’t work, we often blame the LLM or poor prompt engineering. But the real issue is usually the schema. Let me explain: You ask: “Which publishers have received royalty payments above $5,000?” Vector search pulls "publisher" and "royalty_ledger" It misses "vendor_agreement", the bridge table that connects them. So the LLM writes valid SQL that returns nothing. This is why vectors fail on real enterprise schemas. They match names. They don’t discover join paths. Foreign keys already define the path. That’s the gap QueryWeaver is built for. It’s an open-source tool by FalkorDB. QueryWeaver turns your schema into a graph. - Tables are nodes. - Foreign keys are edges. Then it walks the path and includes the bridges automatically. It seamlessly handles multi-hop chains. Tested on the BIRD Benchmark using a superhero database expanded to 60 tables, it resolved a 5-hop query by chaining through: superpower → capability_matrix → stakeholder_registry → resource_requisition → budget_allocation You can also run it locally: 𝗯𝗮𝘀𝗵 𝗶𝗱="𝟲𝗯𝟲𝗽𝗿𝗻" 𝗱𝗼𝗰𝗸𝗲𝗿 𝗿𝘂𝗻 -𝗽 𝟱𝟬𝟬𝟬:𝟱𝟬𝟬𝟬 -𝗶𝘁 𝗳𝗮𝗹𝗸𝗼𝗿𝗱𝗯/𝗾𝘂𝗲𝗿𝘆𝘄𝗲𝗮𝘃𝗲𝗿 I've shared the link to the repo in the replies.

Most “AI agents” forget everything. Why? Because they don’t actually have memory. In this guide we show how to build graph-based agent memory with @mem0ai + @falkordb so agents can remember users, facts, and relationships across sessions. falkordb.com/blog/graph-mem…

Vector search treats every fact as an isolated point. A knowledge graph traverses real relationships between entities. New release: mem0-falkordb: graph-structured memory for agents with per-user isolation, 0 cross-tenant leakage & more: github.com/FalkorDB/mem0-…

If your RAG system costs more to maintain than your LLM inference, you aren't building an AI, you're building a money pit. The "RAG Tax" is real, and vector-only search is the primary collector. Switch to GraphRAG. Cut tokens by 80%. Save your margins.

PST: luma.com/1ewbzub1 GMT: luma.com/m71fs7s7

SAVE THE DATE: Join us for a technical session on advanced graph algorithms, with a cyber-security use case, user-defined-functions, and the recently-launched FLEX library by FalkorDB's CTO and Solution Architecture team. 17 + 18 March, 2026 (two sessions), links below

Don't trust AI agents with your database. They're unreliable, and if you can't see what queries they're running, you can't debug, optimize, or trust them. Most teams building Text2SQL are yolo'ing it. No query validation. No cost tracking on token usage. No fallback when the model hallucinates schema. No guardrails on data access. Building a demo → works great. Production with real enterprise data → total chaos. This is why we built the semantic layer into QueryWeaver. You need to see every query, validate every result, and maintain control over what the agent can actually touch.

Thanks Gartner for recognizing FalkorDB as a leading GraphRAG vendor! If you’re building AI applications, combining Knowledge Graphs with LLMs is today’s smartest strategy. GraphRAG enhances retrieval performance by 35–46%, especially for complex, multihop, and temporal queries

Security data is inherently Graph-shaped. Security data is graph-shaped. Incidents span identity, workload, cloud & network, all highly interconnected. We're coming to present this & more at RSAC 2026. If you'd like to learn more, let's connect #RSAC

FalkorDB's new canvas web component lets you embed the same graph anywhere: dashboards, docs, or your product UI. Install from npm and render on your terms. github.com/FalkorDB/falko…

FalkorDBlite 0.8.0 adds UDFs: custom logic inside graph queries. Easier to standardize similarity scoring (e.g., Jaccard on shared neighbors) and reuse traversal/filter rules across services. Check it out: github.com/FalkorDB/falko…

🚀 FalkorDBlite 0.8.0 is out! This release brings UDF (User-Defined Functions) support, add custom logic directly into your graph queries and make FalkorDBlite even more powerful & flexible. Upgrade, try it out, and tell us what you build! github.com/FalkorDB/falko…

BIG updates coming this month: 1. Context Graphs: Workshop 2. Major announcement on consuming FalkorDB on a major data platform (hint:☃️) fans of data analytics would appreciate this! 3. GraphRAG-SDK enhancements We'd also like to welcome @DvirDu & Guy Lubovitch to the team!!

This Wednesday: Extend your FalkorDB’s functionality with your own set of functions (UDFs); Hands-on workshop + Q&A. January 28 10:00 AM - 11:00 AM PST Sign up: luma.com/lk0stbj7 (Even if you can't attend, we'll send the recording later!)