A fintech team replaced 14 cron-job Python scripts — none of them documented — with visual pipelines. The business team can now read the logic without engineering help.
Finance · Data Pipelines
Finance Data Pipelines
One platform. Every solution. From drag-and-drop to production.
The open-source platform for workflow automation, AI agents, data pipelines, dashboards, and full applications. No-code to pro-code. Free for individuals — forever.
Free forever No account required No-code & pro-code Works offline Open source On-prem ready
Flow-Like Studio
Start with the problem, not the tool.
Solution engineering means you describe what you need — we handle everything else.
You focus on your business logic. Flow-Like manages the rest.
Infrastructure
Desktop, cloud, edge, air-gapped — deploy anywhere with one click.
Security
WASM sandboxing, SBOM, zero telemetry, full audit trail built in.
Observability
Every execution logged. Every AI decision traced. Every cost tracked.
Type Safety
Typed connections between nodes. Errors caught before execution, not after.
Governance
Approval gates, version control, role-based access. Enterprise-ready out of the box.
Scaling
From a Raspberry Pi to a production cluster. Same workflow, same engine.
Solution Engineering in Practice
What teams build with Flow-Like.
Flow-Like isn't a different tool depending on who you are — it's the same workflow automation engine, used differently. Here's what that looks like for real teams.
IT Workflow Automation
Automate onboarding across 5 systems in a day
HR submits a form. Flow-Like provisions accounts in Azure AD, assigns Jira boards, invites to Slack channels, and notifies the manager — all in one visual workflow that anyone on the team can read and modify.
- +Visual workflows replace undocumented scripts — when someone leaves the team, the logic doesn't leave with them
- +Built-in error handling and retry logic, not duct-taped together
- +On-prem or air-gapped deployment if your security team requires it
Good to know
Flow-Like is still in alpha. Enterprise support and SLAs are available, but expect to be an early adopter, not a late majority buyer.
Visual Data Pipeline Builder
Build ETL pipelines your business team can actually read
Pull daily sales from Postgres, clean and validate, aggregate by region, and push to a live BI dashboard. Every connection is type-checked — schema mismatches are caught before the data pipeline runs, not at 3am.
- +Type-safe connections across the entire ETL pipeline — no more "it worked on my machine"
- +Non-engineers can read, annotate, and validate pipeline logic on the same canvas
- +The same data pipeline scales from prototype to production without rewriting in another tool
Good to know
If you need deep integration with dbt, Airflow, or Spark ecosystems, check the node library first. Coverage is broad (1,000+ nodes) but not every niche tool has a node yet.
AI Agent Builder
Ship agents with guardrails and full observability
Build a support agent: user message → PII redaction → Claude with RAG → quality gate → response. Every prompt, token count, latency, and cost is logged automatically. Swap models or add guardrails without touching code.
- +Full execution trace on every LLM call — the observability most agent frameworks bolt on, Flow-Like builds in
- +Model-agnostic: OpenAI, Anthropic, Google, or local via llama.cpp — switch with a node swap
- +Built-in RAG with vector search, PII detection, and approval gates as drag-and-drop canvas nodes
Good to know
If you're deep in the LangChain/LangGraph ecosystem, Flow-Like is a different paradigm (visual DAG vs. code-first). Powerful, but a different mental model.
No-Code Business Automation
Build what you need without filing an engineering ticket
Pull open deals from your CRM every Monday. Have an AI summarize trends and risks. Push a visual report to a Slack channel. All built on the drag-and-drop canvas, no code, no waiting for a sprint slot.
- +Visual, drag-and-drop workflow builder — your domain expertise matters more than coding skill
- +FlowPilot AI assistant helps you build: describe what you want in plain language
- +Same governance and approval gates as engineering-built solutions
Good to know
Complex integrations may still need an engineer to create a custom node. Flow-Like makes collaboration easy, but it doesn't eliminate the need for technical help on edge cases.
Self-Hosted Automation for Makers
Build a smart home dashboard for free, on a Pi
Connect MQTT sensors, transform power data into kWh, detect anomalies, and display everything on a mobile dashboard. Runs locally on a Raspberry Pi. No cloud. No subscription. No account.
- +Free workflow automation — no execution limits, no account wall, works fully offline
- +Runs on a Raspberry Pi, a laptop, or a NAS — not just data center hardware
- +Open source you can fork, extend, and contribute to
Good to know
The community is still growing (200+ Discord members). More general-purpose and less home-automation-specific than Home Assistant — but more powerful when you need to combine IoT with data pipelines, AI agents, or custom UIs.
Internal Tool Builder
Ship apps in hours, not sprints
Build a feature request tracker: custom form UI → AI categorization → manager approval gate → auto-create Jira ticket. The internal tool you built in a morning is the same one running in production six months later.
- +Build complete applications (UI + logic + integrations) in one project, not three tools
- +No "rewrite for prod" phase — the platform enforces architecture from the start
- +Ship to web, desktop (Win/Mac/Linux), or mobile from the same workflow
Good to know
Flow-Like's UI builder shipped in v0.0.8. It's capable but newer than the workflow engine. For pixel-perfect custom UIs, you may still want a designer writing custom components alongside the canvas.
1,000+ Nodes
15+ Languages
~0.6ms Execution
1,000x Faster
0 Rewrite for Prod
"We replaced three separate tools and cut our deployment time by 80%."
"Our business analysts build automations that used to require a full sprint."
"Runs on my Pi. No cloud. No subscription. This is what open source should be."
Solutions You Engineer
Workflow Automation for Every Use Case — BI, AI, Data, Apps.
1,000+ built-in nodes covering the full spectrum. From data pipelines and ETL to customer-facing applications — with infinite room to extend using our developer SDK.
BI Dashboards & Analytics
Interactive dashboards pulling from any data source — real-time, embedded, or standalone. A modern BI alternative built on your workflow engine.
"We replaced three Tableau licenses with one Flow-Like project."
Explore →Workflow Automation & RPA
Replace brittle scripts and RPA bots with visual, error-handled workflows anyone on the team can follow.
"HR onboarding went from 3 hours of manual work to one click."
Explore →AI Agents & LLM Pipelines
Build multi-step AI agent systems on the visual canvas. Every prompt, token, cost, and decision logged with full observability.
"Full observability — we know exactly why the agent made each decision."
Explore →Data Pipelines & ETL
Visual ETL/ELT pipeline builder with type-safe connections. Ingest, transform, catalog — no code required.
"Our data team builds pipelines. Business reads them. No translation needed."
Explore →Internal Tools & Applications
Ship web, desktop, or mobile interfaces alongside your business logic — from the same project. No separate front-end tool.
"We shipped an internal tool in a day that would've taken engineering two sprints."
Explore →API Integrations & Connectors
Connect Kafka, MQTT, ERPs, CRMs, REST APIs, and hundreds more. The integration hub for your entire ecosystem.
"Connected our ERP to Slack to Jira in an afternoon."
BI + Automation + AI Agents + Data Pipelines + Applications = One Solution Engineering Project
Not six tools wired together — one project, one deployment. Build a data pipeline that feeds an AI agent that powers a dashboard inside a customer-facing app. That's solution engineering.
Whitelabel Ready OEM Licensing Embed in Your Product
Solution Engineering for Every Team
One platform for business, engineering & design teams.
Business analysts, citizen developers, engineers, and designers — one shared workspace where everyone contributes to the solution in their own language.
Business & Domain Experts
Define requirements visually. Annotate workflows, set approval gates, validate logic — no code needed. Whether you're a citizen developer or a domain expert, your knowledge drives the solution. FlowPilot AI helps you build what you describe in plain language.
Engineers & Architects
Write custom nodes in 15+ languages. Define type contracts, configure infrastructure, own the execution layer — with the full Rust SDK. Low-code when you want speed, pro-code when you need control.
UX & Interface Designers
Design front-end interfaces wired directly into the workflow automation engine. Ship dashboards, forms, and apps connected to live data from day one.
Zero Rewrite
From prototype to production — without rebuilding anything.
The demo you build on Monday ships on Friday. The workflow engine enforces correct architecture, compliance, and performance from the first node. No rewrite required.
1 EXPLORE
Drag, connect, run. Validate in seconds.
2 REFINE
Add error handling, tests, governance.
3 DEPLOYED LIVE
One click. Same workflow. No translation.
Scroll to explore
"The architecture you ship is the architecture you prototyped. The engine enforces type safety, compliance, and performance from node one. Zero translation layer. Zero rewrite."
The Solution Engineering Engine
A visual workflow engine at its core.
Every solution you engineer — dashboard, AI agent pipeline, or full application — runs on the same type-checked, high-performance Rust workflow engine.
The Visual Workflow Editor
Type-safe drag-and-drop building
Drag nodes onto the canvas, connect them with typed pins, and the visual workflow editor validates every connection before you run. Mismatched data types? Caught instantly — not after your demo crashes.
Built-in version history lets you branch, merge, and roll back. Every change tracked, every collaborator visible. No-code for business users, pro-code for engineers — on the same canvas.
A full type system making sure your solution makes sense before you ever hit run.
FlowPilot
AI Copilot for Workflow Automation
Describe what you need in plain language. FlowPilot, the built-in AI assistant for workflow building, translates your intent into typed, connected workflows — then helps you debug, optimize, and document them.
Context-aware: FlowPilot knows your project's nodes, types, and data sources. Ask it to explain a flow, suggest improvements, or generate new workflow sections.
Less chatbot, more colleague who already read the docs.
The Rust Workflow Engine
~0.6ms execution, 244k events/second
Compiled Rust workflow engine running a typed DAG with Protobuf serialization. File-based projects on an object store abstraction — local filesystem, S3, R2, GCP, or Azure Blob.
~0.6ms per execution. 244k events/second. Your infrastructure costs drop — or you don't need much infrastructure at all.
Custom nodes run in WebAssembly (WASM) sandboxes for isolation and security. The visual editor is your entry point, but there's no ceiling.
→ Architecture deep-dive for CTOs & tech leadsCompiled Rust Typed DAG Protobuf WebAssembly Sandboxing
Trusted in Production
Open-source and self-hosted. Genuinely.
Not "open core" with the interesting parts behind a paywall. The entire codebase is open, governance is built in, and you can inspect everything. Self-host on your infrastructure with full data sovereignty.
0/6 checks passed Done
Access Control
Role-based permissions with approval workflows. Define who can view, edit, approve, and deploy across teams.
Audit Trail
Full execution trace on every workflow run. Inputs, outputs, decision branches, timestamps, user context.
SBOM Published
2,073 dependencies, all tracked. Complete software bill of materials with each release.
Verify →Compliance
SOC 2 aligned, TISAX controls, Prowler assessed. Continuous assessment, not point-in-time.
Verify →Open Source Self-Hosted EU-Based SOC 2 Aligned SBOM Published Prowler Assessed GDPR Ready On-Premise Deployment
Real-World Use Cases
From weekend projects to mission-critical workflow automation.
An enterprise shipped a customer support agent with full observability — every prompt, every token, every cost logged with a complete, auditable decision trace.
Enterprise SaaS · AI Agents
AI Agents Enterprise
A government contractor deployed Flow-Like in an air-gapped environment: zero telemetry, complete SBOM, no modifications required to pass a strict security review.
Government · Air-Gapped Deployment
Government Security
A 2,000-person manufacturer reduced a 3-hour, 5-system onboarding workflow to a single click — built by IT, readable and approvable by operations without translation.
Manufacturing · Process Automation
Manufacturing Automation
A fintech team replaced 14 cron-job Python scripts — none of them documented — with visual pipelines. The business team can now read the logic without engineering help.
Finance · Data Pipelines
Finance Data Pipelines
An enterprise shipped a customer support agent with full observability — every prompt, every token, every cost logged with a complete, auditable decision trace.
Enterprise SaaS · AI Agents
AI Agents Enterprise
A government contractor deployed Flow-Like in an air-gapped environment: zero telemetry, complete SBOM, no modifications required to pass a strict security review.
Government · Air-Gapped Deployment
Government Security
A 2,000-person manufacturer reduced a 3-hour, 5-system onboarding workflow to a single click — built by IT, readable and approvable by operations without translation.
Manufacturing · Process Automation
Manufacturing Automation
A maker runs solar panels, heat pump, and power consumption dashboards on a Raspberry Pi using Flow-Like — fully offline, no subscriptions, no cloud dependency.
Smart Home · Local-First
Makers IoT
FlowPilot turns a plain-English description into a working workflow. Non-technical users tweak nodes and ship automations — no knowledge of typed DAGs required.
Business Automation · No-Code
No-Code Business
A weekend project: a personal finance tracker with AI categorization that runs as both a desktop and a web app — without writing a single line of backend code.
Personal Finance · Cross-Platform
Makers Applications
Freelance developers whitelabel Flow-Like for clients — each gets a branded automation tool, while the developer earns recurring revenue through OEM licensing.
Whitelabel · Developer Revenue
Whitelabel Developer
A maker runs solar panels, heat pump, and power consumption dashboards on a Raspberry Pi using Flow-Like — fully offline, no subscriptions, no cloud dependency.
Smart Home · Local-First
Makers IoT
FlowPilot turns a plain-English description into a working workflow. Non-technical users tweak nodes and ship automations — no knowledge of typed DAGs required.
Business Automation · No-Code
No-Code Business
A weekend project: a personal finance tracker with AI categorization that runs as both a desktop and a web app — without writing a single line of backend code.
Personal Finance · Cross-Platform
Makers Applications
Freelance developers whitelabel Flow-Like for clients — each gets a branded automation tool, while the developer earns recurring revenue through OEM licensing.
Whitelabel · Developer Revenue
Whitelabel Developer
600+ GitHub Stars
200+ Discord Members
100% Open Source
Developer SDK & Platform
1,000+ built-in nodes. Write custom nodes in 15+ languages.
// Define a custom node in Rust — 15 lines use flow_like::prelude::*; #[derive(Node)] pub struct SentimentAnalyzer { #[input] text: String, #[output] score: f64, #[output] label: String, } impl Execute for SentimentAnalyzer { async fn run(&self, ctx: &Context) -> Result<()> { let result = ctx.analyze(&self.text).await?; self.score.set(result.score); self.label.set(result.label); Ok(()) } }
# Define a custom node in Python — 12 lines from flow_like import Node, Input, Output, Context class SentimentAnalyzer(Node): text: Input[str] score: Output[float] label: Output[str] async def run(self, ctx: Context): result = await ctx.analyze(self.text) self.score.set(result.score) self.label.set(result.label)
// Define a custom node in TypeScript — 14 lines import { Node, Input, Output, Context } from '@flow-like/sdk'; export class SentimentAnalyzer extends Node { @Input() text: string; @Output() score: number; @Output() label: string; async run(ctx: Context): Promise<void> { const result = await ctx.analyze(this.text); this.score.set(result.score); this.label.set(result.label); } }
// Define a custom node in Go — 20 lines package nodes import "github.com/tm9657/flow-like-sdk-go" type SentimentAnalyzer struct { Text flowlike.Input[string] Score flowlike.Output[float64] Label flowlike.Output[string] } func (n *SentimentAnalyzer) Run(ctx *flowlike.Context) error { result, err := ctx.Analyze(n.Text.Get()) if err != nil { return err } n.Score.Set(result.Score) n.Label.Set(result.Label) return nil }
Terminal
$ flow-like build --node SentimentAnalyzer
[ok] Compiled to WASM (1.2 KB)
[ok] Type-checked inputs & outputs
[ok] Sandboxed & published to registry
$ _
Rust SDK & WASM Sandboxing High-performance nodes in WebAssembly sandboxes. Full type safety, full isolation.
15+ Language Support Rust, Python, TypeScript, Go, Java, C#, Kotlin, C++, and more. Your language, your nodes.
1,000+ Built-in Nodes Data connectors, AI models, transformations, UI components, protocol adapters.
Open-Source Core Full engine, full SDK, full API. Fork it, embed it, contribute to it.
Multi-Cloud & Edge Deployment
One workflow engine. Every deployment target. Raspberry Pi to production cluster.
Your laptop, a customer's data center, edge computing devices, or any cloud. Offline-first, sync when connected. Multi-cloud deployment with a single workflow definition.
15+ Language Support
Rust, Python, TypeScript, Go, Java, C#, Kotlin, C++, and more. Your language, your nodes.
- Custom nodes in TS, Rust, or Python
- Run locally or on any cloud
- Git-based version control
- Supports 15+ languages and frameworks
Supported languages and frameworks
Runs everywhere
Build once, deploy anywhere — from a single binary to a full cloud cluster.
AWS
Azure
GCP
Docker
Kubernetes
Linux / On-Prem
Journal
Workflow automation insights and product updates.
Flow-Like Alpha v0.0.8 — The UI Building and BI Release
It's here! Our first major release of the year, refining self-hosting capabilities, adding a UI-Builder, Web-App access, and BI features.
Read →Flow-Like Alpha v0.0.7 — Self-Hosting, Storage Freedom & Major UX Upgrades
The self-hosting release. Kubernetes & Docker-Compose setups with observability, multi-cloud blob storage, backend-agnostic desktop apps, and substantial UX, DX, and core improvements.
Read →The 24-Hour Solution - Your Automation, Fast
From "wouldn't it be nice" to "holy cow, it works" in one rotation of the earth.
Read →Flow-Like Alpha v0.0.6 — FlowPilot, Agents, Real-Time Collaboration & Massive Integrations
Our biggest update yet. Introducing FlowPilot, a dedicated Agent Builder, Real-Time Collaboration, and a massive library of new nodes for Microsoft Office, Google Workspace, GitHub, and more.
Read →Dify and Flow-Like (2025): Choose the Right Tool for Your AI Automation Needs
An honest comparison of Dify and Flow-Like—two powerful platforms that excel in different scenarios. Learn when to use Dify's GenAI focus versus Flow-Like's typed automation.
Read →Meet the New Agent Interface: Pin-First, Schema-Free (coming in v0.0.6)
No more manual schemas: define pins on your functions and the agent interface does the rest. Reuse functions for deterministic calls, MCP tools supported, more built-ins coming.
Read →Stay in the loop
Get product updates & engineering insights
No spam — just releases, deep-dives, and the occasional workflow tip. Unsubscribe anytime.
FAQ
Quick answers
Getting Started
Is Flow-Like free? everyone
everyone
Yes. Free for individuals with unlimited local execution — no account required, no execution limits, works fully offline. Pro tiers for teams via SaaS or self-hosting. Enterprise gets fixed pricing per use case, unlimited users, unlimited executions.
Do I need to code to use Flow-Like? business hobbyist
business hobbyist
No. The visual drag-and-drop workflow builder lets you build complete solutions by connecting nodes on a canvas. No-code for business users, pro-code when engineers want to extend. Coding is available but never required.
Can I use Flow-Like without internet? maker enterprise
maker enterprise
Yes. Flow-Like is offline-first. The desktop studio works without any internet connection. Self-hosted deployments can run fully air-gapped. Sync when connected, or never connect at all.
Technical
Can I build AI agents with Flow-Like? ai-builder developer
ai-builder developer
Yes. Flow-Like is a visual AI agent builder with built-in RAG, vector search, PII detection, model-agnostic LLM nodes, and full observability. Build multi-step agent systems with guardrails on the canvas.
Which AI models are supported? ai-builder developer
ai-builder developer
OpenAI, Anthropic Claude, Google Gemini, and local models via llama.cpp — all as drag-and-drop canvas nodes. Bring your own API keys or run fully local. Model-agnostic by design.
Can Flow-Like run fully on-premises? enterprise government
enterprise government
Yes. Designed for on-premise, air-gapped, private cloud, or hybrid deployments. Self-host on your infrastructure — your data never has to leave.
How does Flow-Like handle LLM governance? enterprise ai-builder
enterprise ai-builder
Every LLM call is logged with inputs, outputs, token counts, costs, and timing. Enforce approval gates, rate limits, and model allow-lists at the workflow level. Full audit trail on every AI agent execution.
What does "audit-ready" actually mean? enterprise government
enterprise government
Every workflow execution generates a full trace: inputs, outputs, decision branches, timestamps, and user context. Export in standard formats for compliance teams. SOC 2 aligned, TISAX controls, continuous Prowler assessment.
Comparisons
How does Flow-Like compare to n8n? developer ai-builder
developer ai-builder
Both are open-source workflow automation tools. Flow-Like differs in three ways: a Rust-powered engine (~0.6ms vs. Node.js execution), a full type system that validates connections before runtime, and the ability to build not just automations but also BI dashboards, AI agents, data pipelines, and full applications in one project. n8n is excellent for event-driven automation; Flow-Like is a solution engineering platform for teams building more complex systems.
How does Flow-Like compare to Zapier or Make? business hobbyist
business hobbyist
Zapier and Make are cloud-only SaaS automation tools. Flow-Like is open-source, self-hostable, and runs offline. It also goes beyond automation: you can build BI dashboards, AI agents, data pipelines, and full applications on the same canvas. If you need a free, self-hosted alternative with more power, Flow-Like is worth evaluating.
Can we migrate existing automations? enterprise developer
enterprise developer
Yes. Our onboarding includes migration support for RPA bots, Python scripts, and orchestration tools. We'll audit, document, and convert your highest-value workflows.
Start Building Solutions
What would you build if the platform wasn't the bottleneck?
Download the Studio and start building — locally, offline, no account required. Or jump into the web app. Free workflow automation for individuals, forever.