Vectorize 2.0 - Complete RAG agents (chatbot, MCP) with little or no code
by•
Vectorize 2.0: Your most requested features together:
Chat agents - hosted, no-code chatbots
Chat widget - one-line website integration
Remote MCP - connect to Claude, Cursor, more
Real-time pipelines - always-on syncing
Smarter retrieval - hybrid search, KG
Replies
Vectorize
YouMind
@chris_bartholomew Hi, Chris! Congrats for the launch!
ScaryStories Live
@chris_bartholomew Congrats on shipping Vectorize 2.0! 🎉 One quick thought: in enterprise settings, the real bottleneck isn’t just speed or features, it’s explainability and trust. I’ve seen vector-only RAG pipelines deliver fast results but fall short when stakeholders ask why an answer was produced. Provenance, multi-hop reasoning, and drift prevention become critical here. Have you thought about how Vectorize 2.0 will surface explainable traces or lean further into Graph-augmented RAG to give users not only speed but also confidence in the outputs? That could be a real differentiator as the space gets more crowded!
Vectorize
@tonyabracadabra Thanks, Tony. We are constantly trying to improve the confidence in the outputs of our RAG pipelines. Our chatbot and widget include thinking traces so you can cross check the results. On top of graph-augmented RAG, as you already mentioned, we are launching hybrid search (text + semantic) and advanced filtering capabilities for our autogenerated (by Vectorize Iris, our fine-tuned model) metadata in Vectorize 2.0 which give AI agents powerful tools to get the exact data they need. Plus we have an exciting new capability in the works to improve semantic search on private data. Coming soon!
alphaAI Capital
@chris_bartholomew This is really cool, congrats on the launch!
@chrislatimer @nicoloboschi @jamie_ferguson3 @chris_bartholomew Wow, huge congrats to the team on launching Vectorize 2.0! Seriously impressive how you've made complete RAG agents so accessible with the 'little to no code' approach – that's a game-changer for so many. What really caught my eye is the "smarter retrieval" with hybrid search and knowledge graphs; it sounds like it could lead to incredibly powerful and precise applications. It truly feels like you're democratizing advanced AI, which is awesome to see. As you look ahead, I'm curious: how does the knowledge graph component manage to keep up with dynamic or frequently changing data to maintain top-notch retrieval accuracy?
Vectorize
@aifigureapp Thanks, Jimmy! The biggest challenge with knowledge graph is creating and maintaining the graph data. Vectorize pipelines automatically extract graph data (entities and relationships) whenever there is new or changed data and then automatically writes this to the graph database (ex Neo4j). Our pipelines also do semantic de-duplication on the graph data so there aren't multiple entries with the same meaning. For example, you won't get relationships like WORKS_FOR and EMPLOYED_BY. Those are the same relationship and will be de-duplicated.
With these features we think Vectorize pipelines is the easiest way to set up and maintain graph RAG.
Triforce Todos
Congrats team 👏 Do you already have examples of companies using the real-time pipelines in high-stakes environments?
Vectorize
@abod_rehman Yes, we have B2C AI companies using our real time pipelines. Consumers upload/connect their data and expect to see it incorporated into the answers/results immediately. They don't want to wait a hour or two for the data to get synced. In these cases, our real-time pipeline shine, processing the data almost immediately and making it available to the AI models.
@chris_bartholomew congratulations for your new launch!
Awesome update, can't wait to try out the new remote mcp server!
Hello, your chatbot has an excellent feature. I am amazed with its working style. Thank you so much for your creation.
Vectorize
@ishani_bhattacharya Glad you like it Ishani!
Feels like it could become part of everyday workflow quickly.
Vectorize
@annarobertson We'd love that! Let us know what we can do to make it a reality.
Honestly, looks pretty solid already
Vectorize
@miklesh_pal Thanks! The engineering team has put a ton of effort into it. Glad it shows!
GraphBit
Big release, love how Vectorize combines hybrid search + real-time pipelines with a no-code agent widget. Curious, how does the knowledge graph layer impact retrieval speed under load?
Vectorize
@musa_molla We do combined semantic and knowledge graph retrieval. So you get semantic search results and then related entities. We support Neo4j as the graph database and we find that it performs well under load. The graph results are optional, so depending on your application, you can tune the queries and decide if the graph data is useful on a query-by-query basis.
Love how you bundled the most requested features together—this looks really useful!
Vectorize
@bipul1 Thanks Bipul! We have worked on a lot of AI agents and chatbots so we tried to build what would work best for most people.