Build AI agents and automations by connecting nodes. Run them in your browser —
or deploy to the cloud and they keep running with your laptop closed.
🎁 Try instantly with free daily credits — no API key, no signup. A free account gets 25/day + 100 bonus credits.
Or use your own API keys / local models (Ollama, llama.cpp, vLLM) for unlimited usage.
INTEGRATED WITH
Drag, drop, connect. Code when you want to.
OpenAI, Claude, Gemini, Groq, DeepSeek, xAI (Grok). Switch providers per node. Mix and match for different tasks.
Run Llama, Mistral, Phi, Qwen locally via Ollama, llama.cpp, vLLM, or LM Studio — any OpenAI-compatible server on your network. Private and free. Note: some nodes work best with 32k+ context models.
ReAct-style agents with tool use. Web search, code execution, file access. Chain them together for complex tasks.
Full Python via Pyodide. Pandas, numpy, requests, and more. Runs locally in your browser, not on a server.
AI-controlled web scraping. Navigate sites, extract data, fill forms. Describe what you want in plain English.
Discord, Slack, Telegram, Airtable, Notion, Sheets, Gmail, Twilio — plus bridges into n8n, Zapier, and Make. HTTP nodes connect to any API. If it has an endpoint, you can use it.
Whisper for transcription, Kokoro for text-to-speech. Both run locally. Twilio integration for phone calls.
Embed documents, chunk text, search semantically. Build knowledge bases your AI can actually reference.
Custom Python nodes — describe what you need and AI writes the node for you. Serial and USB HID nodes for hardware. If the browser can do it, you can build a node for it.
Deploy a flow to run server-side on a schedule — watch an inbox, triage, and respond 24/7 with your browser closed. Per-run step logs show exactly what happened while you were away.
Approval gates pause a flow until you say go — get a push notification, review the AI's draft on your phone, then approve or reject. AI drafts it; nothing sends without you.
Package any canvas into a reusable node with its own ports. AI Router sends text down the right path by meaning, and Structured Output turns messy text into validated JSON fields.
Bring your own API keys. We don't mark them up. Or run local models — Ollama, llama.cpp, vLLM — for completely free inference.
Fully Local: Local LLMs (Ollama, llama.cpp, vLLM), Whisper, Python, and local saves stay on your machine. Period.
Browser Mode: When you use cloud LLMs (OpenAI, Claude, etc), those calls go through our server. Not our choice - browsers block direct API calls for security reasons (CORS). We don't log or store your prompts, but if that still makes you uncomfortable, we get it.
Desktop Runner (Pro): Skips the browser entirely. API calls go straight to providers. Nothing touches our servers.
Free daily credits to build. BYOK or local models for unlimited. Pay for extra features.
🎁 Free accounts get 25 credits every day plus a 100-credit signup bonus. Or bring your own API keys / run local models (Ollama, llama.cpp, vLLM) for unlimited usage.
Build Real Workflows
Full Access
Work Together