Feature Description
Add Feedo as a native Vector Store node in Flowise. Feedo is a P2P vector search and storage protocol, a decentralized network of nodes that lets users store and query vectors without self-hosting any vector DB or relying on centralized cloud providers.
Feature Category
Integration
Problem Statement
Users who want RAG in Flowise currently have to either self-host a vector DB (Chroma, LanceDB, Qdrant) or pay for a cloud provider (Pinecone). Both approaches create operational overhead and vendor lock-in:
- Self-hosted DBs require infrastructure, maintenance, and backups.
- Cloud providers are expensive at scale and centralized.
There is no lightweight, serverless option.
Proposed Solution
Add Feedo as a native Vector Store node. Feedo is a P2P vector search and storage network where users store and query vectors across distributed nodes without running local infrastructure.
How it works under the hood:
- Search Nodes store text chunks and corresponding vectors in LanceDB
- Semantic sharding via KMeans centroids - each node stores only its own shard of the embedding space and routes vectors to the correct peer automatically
- Standard nearest-neighbor similarity search (cosine/L2) with low-latency routing
- Raw content lives in a separate Storage Node layer
- Every query is signed by the user's DID key, built-in provenance, no anonymous access
We already have a TypeScript SDK with search and indexing methods, so the integration would be a single Vector Store node class wrapping the SDK. I'm happy to write the PR myself.
Mockups or References
Reference: Feedo GitHub repo - https://github.com/Ashixi/feedo
Similar pattern: existing Pinecone/Chroma nodes in Flowise - ours would follow the same structure.
Additional Context
No response
Feature Description
Add Feedo as a native Vector Store node in Flowise. Feedo is a P2P vector search and storage protocol, a decentralized network of nodes that lets users store and query vectors without self-hosting any vector DB or relying on centralized cloud providers.
Feature Category
Integration
Problem Statement
Users who want RAG in Flowise currently have to either self-host a vector DB (Chroma, LanceDB, Qdrant) or pay for a cloud provider (Pinecone). Both approaches create operational overhead and vendor lock-in:
There is no lightweight, serverless option.
Proposed Solution
Add Feedo as a native Vector Store node. Feedo is a P2P vector search and storage network where users store and query vectors across distributed nodes without running local infrastructure.
How it works under the hood:
We already have a TypeScript SDK with search and indexing methods, so the integration would be a single Vector Store node class wrapping the SDK. I'm happy to write the PR myself.
Mockups or References
Reference: Feedo GitHub repo - https://github.com/Ashixi/feedo
Similar pattern: existing Pinecone/Chroma nodes in Flowise - ours would follow the same structure.
Additional Context
No response