GitHub
Start free
Dashboard
Start freeSign inDashboard

Krypton Bot is Powered by 38 AI providers

Your bots, running on the model you trust

Start on included Krypton AI credits, sign in with the ChatGPT, Claude, Copilot or Grok plan you already pay for, bring a key from 30 providers, or point a bot at your own server. Mix them freely: every bot picks its own model.

Start freeOpen Settings in the app
  • ChatGPT
  • Claude
  • Gemini
  • Grok
  • GitHub Copilot
  • DeepSeek
  • Mistral
  • Meta Llama
  • Qwen
  • Kimi
  • OpenRouter
  • Groq
  • Amazon Bedrock
  • Ollama
  • See all 38
Krypton Cloud · every plan, Free too

Krypton AI: answers from the first minute

No key, no sign-in. Krypton AI sits first in Settings, Models, Providers and is paid from your plan's included credits, so a new bot can work the moment you sign up.

  • Included credits run any OpenRouter model at its real price.
  • Starts on DeepSeek V4.1 Flash, fast and inexpensive, with an automatic fallback.
  • Top-ups never expire. Switch any bot to your own provider whenever you like.

Self-hosting? Krypton AI is a Cloud feature; connect your own providers below.

Already paying for an AI plan? Use it.

Four subscriptions sign in straight from the app. The plan you already have pays for the model; Krypton adds nothing.

  1. ChatGPT Plus or Pro

    Sign in with ChatGPT using a one-time device code.

    Your OpenAI subscription pays for the model. Shown in the app as OpenAI Codex.

  2. Claude Pro or Max

    Sign in with Claude, then paste the code it shows you back into Krypton.

    Your Anthropic plan pays for the model. An Anthropic API key works too.

  3. GitHub Copilot Copilot plan

    Sign in with GitHub Copilot using a one-time device code.

    Runs the Copilot models your plan allows.

  4. Grok SuperGrok or X Premium

    Sign in with SuperGrok or X Premium using a one-time device code.

    Prefer a key? An xAI API key works too.

Bring your own key, 30 providers deep

Paste a key in Settings, Models, Providers and its models appear in every bot's picker. Models a provider refuses are hidden for you, and OpenRouter, DeepSeek and Gemini keys are checked live when you save them.

Model labs

Straight from the companies that train the models.

14

  • OpenAI

    API key

    GPT models on the Responses API.

  • Anthropic

    API key or Claude sign-in

    Sign-in

  • Google Gemini

    API key

  • xAI Grok

    API key or SuperGrok sign-in

    Sign-in

  • Mistral

    API key

  • DeepSeek

    API key

    DeepSeek Flash and V4 Pro.

  • Meta Llama API

    API key

  • Moonshot AI (Kimi)

    API key

    Global and China endpoints.

  • Kimi For Coding

    API key

  • MiniMax

    API key

    Global and China endpoints.

  • Z.AI (GLM)

    API key

    Includes the GLM coding plan.

  • Qwen (Alibaba Token Plan)

    API key

    Team, China and individual plans.

  • Xiaomi MiMo

    API key

    Pay as you go or a token plan.

  • Ant Ling

    API key

Fast inference and open models

Open-weight models served at speed.

7

  • Groq

    API key

  • Cerebras

    API key

  • Together AI

    API key

  • Fireworks AI

    API key

  • Hugging Face

    API key (access token)

    Through the Inference router.

  • NVIDIA NIM

    API key

  • Baseten

    API key

Gateways and routers

One key, hundreds of models behind it.

5

  • OpenRouter

    API key

    The widest catalog in one key.

  • Vercel AI Gateway

    API key

  • Cloudflare AI Gateway

    API key, account ID and gateway ID

  • OpenCode Zen and Go

    API key

  • Radius

    API key

Cloud platforms

Keep model spend on the cloud account you already have.

4

  • Amazon Bedrock

    API key

    Self-hosted servers can use their AWS role instead.

  • Google Vertex AI

    API key or Google Cloud credentials

  • Azure OpenAI

    API key and endpoint

  • Cloudflare Workers AI

    API key and account ID

Run models on hardware you control

Connect any OpenAI-compatible server with a base URL, an optional key and a model ID. There is no Krypton charge for it.

  • Ollama

    Ollama Cloud on any plan, with its built-in preset and your key. A local Ollama on a self-hosted server.

  • LM Studio

    On a self-hosted server, through its OpenAI-compatible endpoint.

  • vLLM

    Any vLLM server: public HTTPS from Krypton Cloud, or your LAN when you self-host.

  • Any OpenAI-compatible server

    LiteLLM, a hosted endpoint or your own proxy. Enter the base URL, an optional key and the model ID.

On Krypton Cloud

Reach any public HTTPS endpoint: Ollama Cloud, a hosted vLLM or LiteLLM, or your own proxy.

When you self-host

Also reach Ollama, LM Studio or vLLM on localhost or your LAN, and let an operator pin local models for everyone.

Self-hosting guide

Voices for calls and voice notes

Speech and transcription have their own providers, set up the same way.

  • ElevenLabsSpeech, transcription and voice cloning.
  • OpenAISpeech and transcription.
  • CartesiaLowest-latency speech.
  • Fish AudioSpeech, cloning and transcription.
  • Self-hosted voiceKokoro, Whisper and other OpenAI-compatible voice servers.

Who pays for the model

Your Krypton plan covers the bots, their computer and memory. The model bill goes to whoever runs the model, never through a Krypton markup.

PathModel bill goes toKrypton fee
Krypton AIYour Krypton plan credits and top-ups, at each model's real priceIncluded
Subscription sign-inOpenAI, Anthropic, GitHub or xAI, on the plan you already haveNone
Your API keyYour provider, billed to you directly at its own pricesNone, no markup
Local or own serverNobody. It runs on hardware you controlNone
Compare plans and creditsBrowse models and prices

Pick a model. Hire a bot.

Start free on Krypton AI, then connect your own providers in Settings, Models, Providers whenever you want.

Start freeOpen the dashboard

Provider questions

Do I need an API key to use Krypton Bot?

Not on Krypton Cloud. Every plan, Free included, starts with Krypton AI credits, so a bot can answer the moment you sign up. Add a key or a subscription sign-in whenever you want your own provider to carry the bill.

Can Krypton use the ChatGPT, Claude, Copilot or Grok plan I already pay for?

Yes. Sign in with ChatGPT Plus or Pro, Claude Pro or Max, GitHub Copilot, or SuperGrok and X Premium in Settings, Models, Providers. The model runs on that subscription and Krypton adds nothing on top.

Does Krypton add a fee when I use my own key?

No. Your own key or subscription uses no Krypton credits and carries no Krypton fee or markup. Your provider bills you directly at its own prices; your Krypton plan only covers the bots, their computer and memory.

Can each bot use a different model?

Yes. Connect as many providers as you like and pick a model per bot, for example a fast cheap model for triage and a frontier model for writing.

Can I run local models?

Yes. A self-hosted server can use Ollama, LM Studio, vLLM or any OpenAI-compatible server on your machine or LAN, at no charge. Krypton Cloud reaches public HTTPS endpoints such as Ollama Cloud or a hosted vLLM.

Product names and logos are trademarks of their owners, shown only to indicate compatibility. Krypton Bot is not affiliated with or endorsed by them.