Garry Tan

Flowise - Build AI agents, visually

Open source Generative AI development platform for building AI agents & LLM workflows.

Add a comment

Replies

Best
Henry Heng

Introducing Flowise 3.0

For the first time, anyone can build AI agents from just an idea. Simply describe what you want to create, and a draft will be generated for you.

Flowise provides modular building blocks for you to build any agentic systems, from simple compositional workflows to autonomous agents. Some common patterns are:

  • Chaining

  • Routing

  • Parallelization

  • Hierarchy

  • Looping

  • Iteration

Each LLM/Agent has the following augmentations to enable flexible multi agents architecture:

  • Memory

  • Retrieval(RAG)

  • Tool

  • Centralized State

  • Structured Output

🙋‍♂️ Human in the Loop:
This is a core aspect of an agentic system, allowing humans to review tasks performed by agents, or to approve/reject tool calls by agents. For example, reviewing actions before sending an email or booking an appointment.

📝 Form Input:
In addition to the standard chat interface, workflows can also be triggered through a form-based interface. This is ideal for workflows that don't require back-and-forth communication between the user and agents.

🔍 Observability:
Gain full visibility into execution traces to debug issues and optimize performance. You can share traces publicly, allowing others to review, approve, or provide feedback.

Flow Validation:
Designing complex workflows can be challenging. Flow validation helps minimize misconfigurations and human errors by automatically checking for common issues.

And the best part? Everything is open source.

$npm install -g flowise
$npx flowise start


Or try it out for free on Flowise Cloud

Aleksandr Heinlaid

@henryheng94 Flowise lets makers build AI agents visually – no-code creativity unlocked. Supported it. Congrats on launch – I'm launching my product today too, feel free to check it out.

Vinod Kiran

@henryheng94  @aleksandr_heinlaid  Thanks for the support. Congratulations on your product launch. We will certainly take a look at your UI Builder.

Aleksandr Heinlaid

@henryheng94  @vinod_kiran wellcome to "draw on screen" too :)

Supa Liu

The potential for community contributions and the flexibility to build truly customized and complex AI solutions is huge. Looking forward to diving into the architecture and seeing what's possible!

Henry Heng

@supa_l Thanks! We have a vibrant community and lots of contributors! https://github.com/FlowiseAI/Flowise

Ong Chung Yau

@supa_l Love hearing that! Can’t wait to see what you build once you dive into the architecture! 🚀

Samir Rashed
Launching soon!

Looks great! Good luck guys 👏

Ong Chung Yau

@samirrashed Thanks so much! 🙌

Henry Heng

@samirrashed thank you, good luck to your coming launch too!

Samir Rashed
Launching soon!

@henryheng94 You're welcome! Waiting for your feedback

I appreciate it

Kamruzzaman Mayed

Flowise is a game-changer for building AI agents—huge respect for keeping it open source! You should definitely list it on Aixyz .Co—it's free and built to spotlight tools like yours in the AI ecosystem.

Henry Heng

@kamruzzaman_mayed Thank you, will definitely look into it!

Kay Kwak

The design is super clean! I'll definitely give it a try. Congrats on the launch!

Henry Heng

@kay_arkain Lots of revisions and iterations behind the new design!

Ong Chung Yau

@kay_arkain Thank you! Really appreciate the kind words, excited to hear what you think after trying it out! 😊

Gianmaria Caltagirone

In my opinion Flowise genuinely feels like the Figma of AI agents—dragging nodes around and watching the prompt logic unfold in real time is both intuitive and oddly fun. Even non-dev teammates (like me) can prototype a lead-qual bot without touching a line of code. 🙏

A few thoughts:

  • Versioning / staging: a Git-style history or branch system would make it safer to iterate once chains are in production.

  • Scale path: any best-practice guide for containerizing Flowise (Docker, K8s) so we can handle bursts of concurrent calls?

  • Model flexibility: a generic “custom endpoint” node would let us drop in home-grown or fine-tuned models alongside OpenAI/HF without hacky workarounds.

Overall, super slick work! 🚀

Vinod Kiran

@gianmaria_caltagirone Appreciate your feedback. We are already working on some of the thoughts you have expressed.

Henry Heng

@gianmaria_caltagirone Thanks for the input!

1.) Versioning is coming up soon

2.) Yes! https://docs.flowiseai.com/configuration/running-flowise-using-queue

3.) It is supported already

Our team comprises both developers and non-technical members. Flowise bridges the gap by providing a platform where everyone can contribute. The visual interface empowers non-technical team members to participate in building and refining AI applications.

Vinod Kiran

@indrajit_kumar3 Glad that your entire team is able to use the platform effectvely. Do take a look at the new agent flow features launched in this version. We are available on the forums / discord if you need any help.

Jun Shen

No-code AI workflows are a game changer! 😄

Xi.Z
Launching soon!

Been exploring Flowise for building AI agents, and it's a game-changer for anyone working with LLMs! The visual development approach makes it incredibly accessible - you can build complex AI workflows without getting lost in code.

As a developer who's worked with various AI frameworks, what stands out about Flowise is how it democratizes AI agent creation. The drag-and-drop interface doesn't sacrifice power for simplicity - you can create sophisticated LLM workflows while visualizing the entire process. Perfect for both prototyping and production.

The open-source nature is a huge plus. The codebase is clean, well-documented, and the community is actively contributing. Since their first launch in May 2023, they've consistently improved the platform, and this latest version shows how much they've matured.

Some standout features:

Visual workflow builder that actually makes sense

Extensive LLM integration options

Easy deployment options

Active community support

Completely free and open-source

The 5/5 rating is well-deserved. Whether you're a seasoned AI developer or just starting with LLMs, Flowise makes the process intuitive while maintaining the flexibility needed for complex use cases. 🔧🤖

Misbah Syed
Have been an early user of Flowise and love the product and the team! The new design and features are 🔥🔥🔥
Henry Heng

@misbahsy Thanks man! Wouldn't be where we are today without your sharing and tutorials on Flowise!

Ong Chung Yau

@misbahsy Thank you so much for sticking with us since the early days! 🙌

Bryce York

Looks great guys! I'm a huge fan of no-code flow-style tools.

Ong Chung Yau

@bryce_york_ag Thanks a lot! Glad to hear you're a fan of no-code flow tools, hope you enjoy using Flowise! 🙌

David V. Kimball

Looks awesome. Good luck!

Jason Chernofsky

looks awesome! is this to be used instead of or in tandem with n8n?

Henry Heng

@jason_chernofsky We don't have as much integrations as n8n does, but technically with MCP tools, you can achieve the same thing as well. Few things that differentiate us from other automation platforms:

  • Support streaming

  • Multi agents communication (agents have built in memory that can see the full context)

  • Shared state

  • Memory optimization, tool calling, structured output, RAG

  • Human in the Loop

Agentflow is Agentic-first, meaning it is built for designing single/multi agents architecture and workflows.

Paul Plessing

We're using Flowise since more than a year, it's really useful!
Looking forward to see the changes in Agents Flow v2.

If you ever wanted to know how to self host Flowise, here's a video (1 year old but not outdated at all!):

umar naseer

Looks promising — I’ll explore it further. Based on what I’ve seen so far, it seems like the best product on the list

Ong Chung Yau

@umar_naseer1 Thank you! Really glad to hear that!

LBdata

I was wondering if this is the same Flowise we're using to run our RAG and yup it sure is! We handle the training data aspect, let's connect I think your platform is awesome.