Flowise
Open-source AI agent builder — wound down by its team in August 2026.
Overview
Flowise is an open-source, Apache-2.0-licensed platform for building AI agents and RAG pipelines on a visual canvas. Workday acquired the company in August 2025; the team announced a wind-down on 29 July 2026 and the project reached end of life on 31 August 2026. The code remains public and forkable, but no further releases are expected from the original maintainers.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 2 Sep 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- 2 flows & assistants
- 100 predictions/mo
- 5MB storage
- Unlimited flows & assistants
- 10,000 predictions/mo
- 1GB storage
- 50,000 predictions/mo
- 10GB storage
- Unlimited workspaces
- +$15/extra user
- SSO & RBAC
- Audit logs
- 99.99% SLA
- Dedicated support
Use cases
What you can produce with Flowise
- A drag-and-drop canvas for wiring LLMs, vector stores, tools and memory into a single-agent chat flow or a multi-agent "Agentflow"
- Fast prototyping of RAG and agent setups, which is what it was genuinely good at
- An Apache 2.0 self-hosted core — not MIT, whatever some listings still claim, which matters if your legal team checks
- ⚠️ END OF LIFE. The founder froze the code on 29 July 2026, the repository was archived on 13 August, and the project ended on 31 August 2026. Workday's 2025 acquisition funded the project; it did not commit to it
- ⚠️ A remote-code-execution history with no one left to patch it. Three serious RCE vulnerabilities were disclosed inside about a year, one a maximum-severity CVSS 10.0 flaw in the CustomMCP node allowing arbitrary JavaScript execution, with researchers finding over 12,000 exposed instances online. Fixes shipped in v3.0.6/3.1.1, so a current install is covered against those — but the repository is archived, so the NEXT disclosure has no upstream patch coming
- ⚠️ If you run it anyway: you own the fixes yourself, and do not expose the API publicly
- The hosted tiers while they last: free at 2 flows and 100 predictions a month, Starter $35/month for unlimited flows and 10,000 predictions, Pro $65/month for 50,000, Enterprise by quote for on-prem or air-gapped with SSO and audit logs
Flowise reached end of life on 31 August 2026, and that is now the first thing to know about it. Its founder announced the wind-down on 29 July, the repository was archived read-only on 13 August, and npm packages and Docker images were marked deprecated at the same time. The stated reason is a shift in how people build — coding agents absorbing work that rigid low-code canvases handled — not a security failure and not the Workday acquisition.
What it was, and still is if you fork it: a low-code platform for building AI agents and RAG pipelines visually, chaining models, vector stores, tools and memory on a drag-and-drop canvas, built on LangChain and LlamaIndex. Self-hosting stays free under Apache 2.0, the code remains public, and the team explicitly encouraged teams to fork it and maintain their own builds.
The security position has changed in kind, not just in degree. Several critical remote-code-execution flaws were disclosed across 2025 and 2026, one rated CVSS 10.0, with thousands of exposed self-hosted instances found under active attack before fixes landed. Those particular fixes shipped, so a current installation is patched against them. What is gone is the next fix: with the repository archived, any vulnerability disclosed from now on is yours to patch, in your own fork, or not at all. Do not expose the API publicly, and treat a self-hosted deployment as software you have adopted rather than software you are running.
Beyond that: complex flows get hard to debug, and there is no built-in RBAC or audit logging outside the paid tiers — which are themselves winding down. For a new project, the question is no longer whether Flowise fits, but whether you are prepared to maintain it.
What people say
Workday bought Flowise in August 2025, and for a while the open-source repo kept shipping — v3.1.3 landed in June 2026. That is where this entry used to stop, and it read as reassurance about long-term support. It did not hold: on 29 July 2026 the founder announced a code freeze, the repository was archived on 13 August, and the project reached end of life on 31 August. The last push was 13 August 2026, at roughly 55,400 stars. An acquisition funding an open-source project is not the same as an acquisition committing to it.
What you get is a drag-and-drop canvas for wiring together LLMs, vector stores, tools and memory into either a single-agent chat flow or a multi-agent "Agentflow." It's genuinely fast for prototyping RAG and agent setups, and the self-hosted core is Apache 2.0 — free, not MIT as some listings still claim, which matters if your legal team checks these things. Enterprise names on the site now include AWS, Priceline, Accenture and Deloitte.
The uncomfortable fact worth knowing before you self-host: Flowise had three serious remote-code-execution vulnerabilities disclosed within about a year, one of them a maximum-severity CVSS 10.0 flaw in the CustomMCP node that let attackers run arbitrary JavaScript. Researchers found over 12,000 exposed instances online before fixes in v3.0.6/3.1.1 shipped. Those fixes exist, so a current install is covered against them — but the repository is now archived, so the next disclosure has no upstream patch coming. Running this after end of life means owning the fix yourself. Do not expose the API publicly.
Pricing for the hosted option: free for 2 flows and 100 predictions a month, Starter at $35/month for unlimited flows and 10,000 predictions, Pro at $65/month for 50,000 predictions, Enterprise by quote for on-prem or air-gapped deployment with SSO and audit logs. Good fit for technical teams comfortable owning their stack; less so for anyone wanting a fully managed, hands-off service.
Summary of public user & expert reviews, compiled by RECATOOLS.
Notable facts
- Flowise was created by two developers from Singapore, making it a notable APAC contribution to the open-source AI tooling ecosystem.
- The project went from 0 to 20,000 GitHub stars in its first 3 months, driven by developers frustrated with the complexity of building LangChain applications in pure code.
- Flowise can be deployed on any cloud provider in under 5 minutes using the official Docker image.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including Flowise's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Flowise unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Flowise directly →
Spotted something out of date? Suggest an update →
More in Agents & Automation
