FlowLane

Visual AI workflow builder — drag, drop, deploy.

Build AI workflows by connecting nodes visually. Export to SKILL.md for any agent.

Features

• Drag & Drop — Visual node builder with 7 node types
• AI Nodes — Input, AI Prompt, Transform, Condition, Output, API Call, Skill
• SKILL.md Export — Generate structured skill files for Claude, Cursor, Codra
• CLI — Initialize, dev server, build from command line
• MCP Server — Tool calls for workflow generation
• SDK — npm and Python clients

Install

npm install -g @talocode/flowlane

pip install talocode-flowlane

Quickstart

flowlane init
flowlane dev
# Open http://localhost:3000
# Drag nodes, connect them, generate SKILL.md

CLI Commands

flowlane init      # Initialize project
flowlane dev       # Start visual builder (default: port 3000)
flowlane build     # Generate SKILL.md from workflow

Node Types

| Node | Description | |------|-------------| | Input | User input or text | | AI Prompt | Send to AI model | | Transform | Process data | | Condition | If/else logic | | Output | Display or save | | API Call | External API | | Skill | Import SKILL.md |

MCP Server

flowlane-mcp

Tools:

• flowlane_create_workflow — Create a new workflow
• flowlane_generate_skill — Generate SKILL.md from workflow
• flowlane_validate_workflow — Validate a workflow
• flowlane_list_templates — List available templates

API

const { FlowLane } = require('@talocode/flowlane');

const client = new FlowLane({ apiKey: 'your-key' });

await client.createWorkflow({ name: 'chatbot', nodes: [...], edges: [...] });
await client.generateSkill({ workflow, format: 'claude' });

Python SDK

from flowlane import FlowLane

client = FlowLane(api_key="your-key")

client.create_workflow(name="chatbot", nodes=[...], edges=[...])
client.generate_skill(workflow=workflow, format="claude")

Links

• GitHub: https://github.com/talocode/flowlane
• X: https://x.com/talocode
• Telegram: https://t.me/TalocodeChannel
• Site: https://talocode.site

License

MIT © Talocode