gkorland

If you want to read more about it check https://www.falkordb.com/blog/beyond-in-memory-graphs/

ActiveGraph.ai v1.2.0 is live * Graph projection is now pluggable

falkordb = native edges + cypher pushdown * 2-hop query: ~8.9s → ~300ms * community-built core arc

graphify is a great project you should try it with FalkorDB.

``` /graphify ./raw --falkordb # generate cypher.txt for FalkorDB /graphify ./raw --falkordb-push falkordb://localhost:6379 ```

Where do you store the Knowledge Graph? Are you supporting any Graph Database like FalkorDB?

Did you consider using a Graph Database? e.g. FalkorDB? If you want to keep it lite we released a lite version that can run embedded in your nodejs.

That’s exactly why we released FalkorDBLite,an embedded graph engine designed for local, edge, and GenAI use cases (GraphRAG, agentic memory, semantic layers, code graphs, etc.).

https://www.falkordb.com/blog/falkordblite-embedded-python-g...

Nice! Do you plan to add support for other Graph Database (e.g. FalkorDB)?

We're building an open-source Text2SQL tool that transforms natural language into SQL using graph-powered schema understanding. Allowing you to ask your database questions in plain English, QueryWeaver handles the "weaving".

https://github.com/FalkorDB/QueryWeaver/

can you share a link?

The fact that GQL is now supported by some of the relational Database, doesn't mean they'll become an alternative to native Graph Databases.

You should check FalkorDB https://github.com/falkordb/falkordb

Did you try to run redis-benchmark? (Compared to Redis)

We built QueryWeaver exactly to solve those pain points—SQL generation that actually understands your business context and keeps the conversation alive across follow-ups. The graph layer is fully extensible, so you can define things like what an “active user” means for your org. Would love to hear your feedback if you give it a try!

You can find the release notes here: https://github.com/FalkorDB/FalkorDB/releases/tag/v4.10.0

Much more than that, FalkorDB added many features on top of RedisGraph.

Your graph DB frustrations mirror what many experienced with Neo4j. If you refresh your project, consider including FalkorDB (formerly RedisGraph) - it uses sparse adjacency matrices and GraphBLAS for much better performance while supporting Cypher.

Would be interesting to see updated benchmarks comparing these newer options against PostgreSQL extensions.

The QirK paper is super interesting https://arxiv.org/pdf/2408.07494.

We took that one step further with the GraphRAG-SDK - https://github.com/FalkorDB/GraphRAG-SDK

It practically supports every models out there, LiteLLM supports over 100 large language model services, including OpenAI, Claude, Gemini, WatsonsX, Mistral, Azure OpenAI, Sagemaker....

We're excited about this version which add support for LiteLLM https://github.com/BerriAI/litellm.

With LiteLLM the application can easily switch between different models to pick to most optimized model for Knowledge Graph generation and Knowledge Retrieval.

We see three major issues that different projects encounter:

1. Knowledge Graph quality - if you don't have a clean well defined Knowledge Graph then the end result will not be good.

2. Multi Graphs support - you want to break the large Knowledge Graph to small per domain Knowledge Graphs which really helps the LLM work with the data.

3. User/Agent memories - You want each user have a dedicated & effective long term memory AKA personal Knowledge Graph completely private and secured.

4. Latency/performance - you have to have a low latency Graph Database that can provide a good user experience.

If it's your personal assassinate and is helping you for months it means pretty fast it will start forget the details and only have a vogue view of you and your preferences. So instead of being you personal assassinate it practically cluster your personality and give you general help with no reliance on real historical data.

Did you try FalkorDB? If latency is a factor and you consider to store many Knowledge Graphs you should check it out.

It looks great! Utilizing Knowledge Graph to store long term memory is probably the most accurate solution compared to using only Vector Store (same as with GraphRAG vs Vector RAG). I think an important thing to point here that long term memory vs RAG doesn't represent the organizational knowledge but the chat history which should be private to the end user and shouldn't be kept in a completely isolated graph than the rest of the users

When using a graph database you can build a knowledge Graph out of the long term memory. Storing it only in a vector database means that you'll only find things that are similar to the user question and miss a lot of information that is an aggregation of different memories.

It's a very similar reason on why using Graph for RAG can help you get much more accurate responses than with Vector RAG. See a blog I wrote about it https://www.falkordb.com/blog/knowledge-graph-vs-vector-data...

Do you just hold this number of node in the database or also need to visualize them all in one view?

We've done a very similar procedure just with FalkorDB as a Graph Database. Notice if you already have a schema/ontology it might be easier might you might miss some entities in the text you did realize exist.

So in our in GraphRAG-SDK we are running two phases, the first is sampling the data to suggest a schema and the second is using this schema to ground the LLM to this schema (as you suggested)

Check https://github.com/FalkorDB/GraphRAG-SDK

Would love to hear your feedback once you do try GraphRAG.

Nice demo! It so easy to see that Vector RAG just can't answer complex queries that are more that Sematic Search results.

We also build a similar demo based on the UFC dataset - https://github.com/FalkorDB/ufc

Nice!

That is cool! Would love to see how it works with FalkorDB as the Graph Database for the knowledge Graph https://github.com/FalkorDB/FalkorDB