Flowise - UI for LangChain
-
Be good to see @cloudron_io on these mentions:
AI tools are vast amounts of free publicity right now.
-
T timconsidine referenced this topic on
-
THIS! Sooo much YES!
+1 -
I tried making a package for Cloudron.
Despite it looking really easy, it isn't.
I tried a local installation on desktop, but it didn't work.
I then tried a docker-compose installation on another VPS.
This worked but documentation on usage is almost non-existent.
I got it working with a simple document.
Key points to note are :- dimensions for the index must be 1536 (have to search github closed issues to discover this)
- in Pinecone retriever, the index has to be the 'simple' name, not the full URL
- also in Pinecone retriever, namespace is what Pinecone calls a Project
Although it worked with a simple document, I can't get a more complex document loaded.
Does Flowise make it easy ?
well, easier than programming the setup manually, yes.
Easy to use ? Hmmmm.Indie app dev, scratching my itches, lover of Cloudron PaaS
-
I tried making a package for Cloudron.
Despite it looking really easy, it isn't.
I tried a local installation on desktop, but it didn't work.
I then tried a docker-compose installation on another VPS.
This worked but documentation on usage is almost non-existent.
I got it working with a simple document.
Key points to note are :- dimensions for the index must be 1536 (have to search github closed issues to discover this)
- in Pinecone retriever, the index has to be the 'simple' name, not the full URL
- also in Pinecone retriever, namespace is what Pinecone calls a Project
Although it worked with a simple document, I can't get a more complex document loaded.
Does Flowise make it easy ?
well, easier than programming the setup manually, yes.
Easy to use ? Hmmmm.@timconsidine I also tried with local Docker and got it running smoothly, BUT the usage was complicated for me. Even the marketplace Templates didn't work well. I like the UI and the Idea, but I was not able to "use" it.
-
M marcusquinn referenced this topic on
-
-
(Thanks for posting screenshots as well since Twitter is now login-walled…
) -
I think this is a safe bet for packaging, seems to be the market leader for AI integrations, and doesn't do anything AI itself, so no GPU needs.
-
this is the docker file: https://docs.flowiseai.com/configuration/deployment/hugging-face
-
Flowise would be a good strategic win, as they are well funded, and AI tool agnostic.
-
It seems they've simplified the install quite a bit - https://github.com/FlowiseAI/Flowise/blob/main/Dockerfile
Their new 3.0 has capabilities that eclipse n8n in state awareness across nodes and simplification.
node app with chromium-browser needed for packaging.
https://github.com/FlowiseAI/Flowise/blob/main/Dockerfile -
A small suggestion for formatting your wish list:
--- ## **Title**: Flowise on Cloudron - Build AI Agents Visually *** * **Main Page**: https://flowiseai.com * **Git**: https://github.com/FlowiseAI/Flowise * **Licence**: Apache License Version 2.0 * **Dockerfile**: Yes * **Demo**: https://flowiseai.com (Free tier available with 2 flows) *** * **Summary**: Flowise is an open-source generative AI development platform that enables building AI agents and LLM (Large Language Model) applications visually, without coding. The tool offers a drag-and-drop interface to create custom workflows, integration with various LLM providers (OpenAI, Anthropic, Azure, etc.), RAG (Retrieval-Augmented Generation) support, conversational memory management, and multi-tool agent capabilities. It features a modular architecture with Node.js backend, React frontend, and third-party integration components. *** * **Notes**: Flowise is an excellent alternative for those looking to implement AI solutions without complex programming. It offers advanced features such as Document Stores, vector databases (Pinecone, Qdrant, Chroma, etc.), response streaming, and monitoring with OpenTelemetry/Prometheus/Grafana. Supports deployment via Docker, Docker Compose, and includes Redis queue configuration for scalability. The community is active with over 21 contributors to the main repository. Concerns include managing credentials for external APIs and the need for proper storage configuration for data persistence. *** * **Alternative to / Libhunt link**: https://www.libhunt.com/r/Flowise * **Screenshots**: - Official logo: https://github.com/FlowiseAI/Flowise/blob/main/images/flowise.png - Demo GIF: https://github.com/FlowiseAI/Flowise/blob/main/images/flowise_agentflow.gif - Additional screenshots available in the GitHub repository and official documentation *** **Additional Technical Information:** - Default port: 3000 - Supports SQLite, MySQL, PostgreSQL as database - Storage: Local, AWS S3, Google Cloud Storage - Requires Node.js 18.15.0 or higher - Available on Docker Hub as official image
