On September 17, 2026, n8n shipped v2.36 with a feature that eliminates the single biggest friction point in automation experimentation: the signup wall. Gateway credits let you use supported AI models and tool services directly inside n8n Cloud without creating accounts, generating API keys, or entering payment methods on a single third-party site. You pick a node, select Gateway credits instead of your own credential, and the workflow runs.
That is a small sentence and a big deal. Anyone who has tried to build a multi-service AI workflow knows the drill. You want to test a template. You add a node for Anthropic. You need an API key. You create an Anthropic account. You add a credit card. You generate the key. You paste it back into n8n. You add a search node. You need a Brave API key. You create a Brave account. Twenty minutes later you still have not run the workflow, and your enthusiasm has quietly evaporated. Gateway credits compress that entire detour into a single dropdown selection.
What six model providers actually means
The supported models at launch are Anthropic, Google Gemini, OpenAI, Alibaba Cloud (Qwen), Kimi.ai, and MiniMax. That mix is worth pausing on. The first three are the providers most automation users already know. The last three are the ones most have been curious about but never bothered to set up.
Qwen is Alibaba's open-source model family, positioned for coding agents and long-horizon enterprise tasks. Kimi.ai is Moonshot's model with both thinking and instant modes, designed for multi-step research and coding. MiniMax offers M3 for agentic work across a million-token context and H3 for omni-modal video generation. These are not toy models bolted on for marketing purposes. They are serious contenders that most small teams would never trial if it meant maintaining three more vendor relationships.
Gateway credits change the calculus. You can now put Qwen, Kimi, and MiniMax into the same workflow as Claude and GPT, run them side by side, and see which one actually performs best for your specific use case — all from a single prepaid balance without a single extra account.
The tool services matter just as much
The five tool services available at launch are Brave Search, Browserbase, Firecrawl, LlamaParse, and PDF.co. Each solves a different piece of the data pipeline problem that AI workflows run into constantly.
Brave Search gives your workflow a real-time web index that is independent of Google. Browserbase opens a managed cloud browser that can navigate, click, and extract data from pages that require interaction. Firecrawl turns entire websites into LLM-ready data, handling JavaScript rendering and rate limits automatically. LlamaParse does agentic OCR on complex PDFs and documents, preserving tables and layouts that basic extractors destroy. PDF.co handles generation, conversion, merging, and splitting.
Individually, each of these is useful. Together, they cover the full journey from "I need data from the web" to "I need structured output my model can actually work with." And you can try all of them without signing up for anything.
The pricing is deliberately boring
n8n says Gateway credits follow each provider's published list price. There is no premium for the convenience of skipping signup. Your balance shows in real currency, and the Cloud Admin Panel breaks down usage by workflow and by service. The pricing page at app.n8n.cloud/service-pricing lists every rate before you run anything.
This is the right call. The moment you add a mystery markup to a convenience layer, users start doing mental math about whether the time saved justifies the extra cost. By keeping pricing transparent and at list price, n8n removes that friction too. The only thing you are paying for is the model call itself.
One privacy note worth knowing: when you use Gateway credits, the provider receives your request content (the prompt or document) but not your identity or n8n account details. That is a reasonable middle ground. Your data goes to the model because it has to. Your identity does not.
What this means for your next project
If you have been sitting on an automation idea because setting up five API keys felt like a chore, that excuse is gone. Gateway credits turn n8n Cloud into a sandbox where you can prototype multi-model workflows in minutes instead of afternoons. Start with a template or describe what you want to n8n Assistant in natural language. It will apply Gateway credits automatically for any supported service you have not configured a credential for.
The list of supported providers is going to grow. n8n says these eleven are "where we're starting, not where we're stopping." For now, the launch lineup covers enough ground that most automation experiments can get off the ground without a single signup detour. The hardest part of building something useful should be the idea, not the onboarding. Gateway credits finally make that true.
Sources: n8n blog — Try new models and services, skip the account setup (September 17, 2026). n8n Gateway credits documentation.
Created by macbook-air-telegram via Telegram on 2026-09-17