AI ProductsOverview

AI Products

Build and monetize AI-powered products on top of Aerostack’s gateway infrastructure. Drop in the chat widget or use the SDK to build your own UI — Aerostack handles auth, billing, rate limiting, streaming, and usage analytics.

Two ways to deploy

Case 1 — Your own model (BYOC)

Running a fine-tuned model on AWS, Azure, a GPU server, or any custom endpoint? Aerostack puts a production gateway in front of it. No changes required on your server — just expose a standard POST /v1/chat/completions endpoint.

Your users → Aerostack Gateway → Your model server
              ├─ Auth & session
              ├─ Quota + rate limiting
              └─ SSE stream → back to user

Case 2 — AI provider API (BYOK)

Using OpenAI, Anthropic, Gemini, or Workers AI? Bring your API key and optionally compose a pipeline — RAG retrieval, pre/post-processing hooks, model fallbacks — before calls reach the provider.

Your users → Aerostack Gateway → Pipeline → OpenAI / Anthropic / Gemini
              ├─ Auth & session       ├─ RAG (your docs)
              ├─ Quota + rate limit   ├─ Pre-hook (edge fn)
              └─ SSE stream ←─────── ├─ LLM call
                                      └─ Post-hook (edge fn)

What you get

UI options

You have full control over how users interact with your AI product.

OptionPackageWhen to use
Chat widget@aerostack/chat-widgetZero-config embed. Ship in 2 minutes.
React hooks@aerostack/reactCustom UI. Full control over design.
Universal SDK@aerostack/sdkVue, Svelte, React Native, Node.js — any runtime.

Auth is optional. If you already handle auth in your app, use consumer-key-only mode or pass your own JWTs. Aerostack never touches your user data.

Next steps