Frequently asked questions

Three things in one platform: (1) an AI gateway API that unites 200+ models (OpenAI, Anthropic, Google, Mistral, Llama …) behind a single OpenAI-compatible endpoint. (2) A visual pipeline editor that lets you chain multiple models into your own REST API. (3) A reseller widget that lets you sell tokens to your end customers — you keep 95% of every sale.

A pipeline is a chain of AI calls that you build visually. Example: image in → GPT-4 describes it → Claude translates the description → output as JSON. The whole thing automatically becomes a REST API: POST /v1/pipelines/your-slug/run. No servers, no deployment.

No for the editor and the resell widget — both are drag-and-drop. Yes for using the API itself (you need to be able to call fetch()), but we provide copy-paste snippets for JavaScript, Python and cURL.

OpenAI gives you one model. We give you 200+ models behind one API + a visual pipeline builder + a reseller layer. You can replace your OpenAI account with ours instantly — same code, same response structure — and immediately use Claude/Gemini/Llama too, without new accounts.

No. Pure pay-as-you-go. You top up credits and use them whenever you want. Credits never expire.

We pass on the provider prices with a fair markup — the exact margin varies by model, but is never hidden. In the editor you see live what your pipeline costs per run. Free-tier models like Llama-3-8B are even usable for 0 credits (the platform covers the cost).

Yes — 10,000 free credits right after sign-up, no credit card required. Typically enough for a few thousand small model calls.

We log metadata (model, token counts, latency) for billing. Prompts and outputs only end up in the log if your x-log-content header is set — otherwise they're discarded after being forwarded to the provider. Pipelines store their inputs/outputs as pipeline_runs so you can debug — you can delete those in your dashboard.

Frankfurt (Vercel/Neon EU region). Provider calls go to the respective provider — you can use a pipeline setting to restrict to EU-region models (Mistral, EU-OpenAI).

We hash them with Argon2id and store only the hash. You see the plaintext exactly once at creation. If lost → generate a new key, revoke the old one.

You create token packages (e.g. "1000 tokens for €5"), copy a <script> snippet into your site, done. End customers pay you directly via Stripe. We webhook your server that user XYZ now has 1000 tokens — you credit them in your app. We take 5%, you 95%.

No. The Stripe checkout shows only getmorecredits.com — all redirects go through us. Your site stays invisible to Stripe.

The default is weekly; you can change it in your Stripe Connect dashboard.