Works with your favorite AI tools
Deploy AI-powered apps, chatbots, agents, and API wrappers on our standard Web Hosting plans — free tier included. Bring your OpenAI, Anthropic, Replicate, or Bedrock keys and ship in minutes.
View Hosting PlansMost AI projects today are a Next.js or Node app calling a hosted model API. Katika is purpose-built for exactly that.
Run a full Next.js app server-side on the free plan. Cold-start auto-suspend keeps it efficient; your AI routes wake on demand.
OpenAI, Anthropic, Google Gemini, xAI Grok, Replicate, Together, Groq, Bedrock — if it has an API, you can call it from a KWS-hosted app.
Connect GitHub and we redeploy on every push. Or use kws deploy from the CLI. Auto-detects Next.js, Vite, static, and more.
Every plan has hard monthly caps. When you hit one, the site pauses until next month. We literally cannot bill you over your plan price.
Every project gets yourapp.katikaws.com with free SSL. Bring a custom domain on any paid plan.
Per-project metering of bandwidth, SSR compute, and storage with email alerts at 50%, 80%, 95%, and 100% of your caps.
Pick the path that matches your app architecture — most don’t need a dedicated GPU.
Your app calls OpenAI, Anthropic, Replicate, Bedrock, Together, Groq, or any other hosted model API. KWS handles the web layer; the AI provider handles inference.
You need to run Llama, Mistral, Stable Diffusion, or a custom fine-tuned model on dedicated GPU hardware. Will be sold as a GPU add-on (shared T4, dedicated T4, or A100) layered on top of any Web Hosting plan.
If you’re an existing AI Hosting customer on the old plan structure, your subscription continues uninterrupted — we’re grandfathering all current customers. Contact support@katikaws.com with any questions.
All of these run great on a standard Web Hosting plan — no GPU required.
Free covers most indie AI apps. Move to Starter, Pro, or Business as your traffic and SSR needs grow.
kws deployPush your Next.js, Vite, or static project. We detect the framework, install, build, and serve.
Set OPENAI_API_KEY, ANTHROPIC_API_KEY, etc. as project env vars. Your app is live at a custom subdomain with SSL.
No. AI projects on Katika run on the same Web Hosting plans as everything else — Free, Starter, Pro, and Business. Most AI apps today are a Next.js or Node app that proxies a hosted model API (OpenAI, Anthropic, Replicate, Bedrock, etc.), and those workloads fit comfortably on standard plans.
Any provider with an HTTPS API. We have first-class examples for OpenAI (ChatGPT, GPT-4), Anthropic (Claude), Google (Gemini), xAI (Grok), Replicate, Together, Groq, Hugging Face Inference Endpoints, and AWS Bedrock. You set the API key as a project environment variable in the dashboard.
Self-hosted GPU inference (running Llama, Mistral, Stable Diffusion, or a fine-tuned model on dedicated GPU hardware) is on our roadmap as an add-on layered onto any Web Hosting plan. It is not generally available yet. If you need GPU hosting today, email support and we will keep you posted.
Every KWS plan has hard monthly caps on bandwidth, SSR compute, and storage. When you hit one, the site pauses until next month or until you raise the cap. The hosting bill cannot exceed the plan price. Your AI provider (OpenAI, Anthropic, etc.) bills you separately for inference, and you control those keys directly.
Yes. You can change plans at any time from the dashboard. Upgrades take effect immediately and billing is prorated.
Works with your favorite AI tools