BYOK apps by AI provider

Start from the key you already have. Each page lists the apps that accept it and explains where to create one.

• OpenAI85 toolsAn OpenAI API key gives apps access to GPT models, image generation, speech-to-text (Whisper) and text-to-speech, billed per token to your own OpenAI account. API usage is billed separately from a ChatGPT Plus or Pro subscription, so a BYOK app lets you pay only for what you use instead of another monthly plan.
• Anthropic64 toolsAn Anthropic API key lets apps use Claude models (Opus, Sonnet and Haiku) at Anthropic's per-token API prices. A Claude Pro or Max subscription does not include API usage, so apps that accept your own key are the way to use Claude inside other tools without paying a reseller's markup.
• Google Gemini62 toolsA Gemini API key from Google AI Studio gives apps access to Gemini models for chat, long documents, images and audio. Google offers a free tier with rate limits, which makes Gemini a common first key to try with BYOK apps; paid usage is billed through Google Cloud.
• OpenRouter31 toolsOpenRouter is a single API and a single key for hundreds of models from OpenAI, Anthropic, Google, Meta, Mistral, DeepSeek and others. Apps that accept an OpenRouter key let you switch models without managing a key per provider, and you pay OpenRouter for usage from one balance. OpenRouter also has its own BYOK option, where you attach provider keys to your OpenRouter account and requests are billed to those keys.
• Mistral29 toolsA Mistral API key gives access to Mistral AI models, including Mistral Large, Mistral Small, Codestral for code and Pixtral for images. Mistral is a European provider, which matters to teams that prefer EU-based processing.
• Ollama29 toolsOllama runs open models (Llama, Qwen, Gemma, Mistral, DeepSeek and others) on your own computer. Apps that support Ollama connect to its local server, usually at http://localhost:11434, so you need no API key, pay nothing per token, and your prompts never leave your machine.
• DeepSeek23 toolsA DeepSeek API key gives access to DeepSeek models, including the reasoning model, at some of the lowest per-token prices among major providers. The API is OpenAI-compatible, so many apps support it as a custom endpoint as well.
• xAI18 toolsAn xAI API key gives apps access to Grok models. Grok on X or in the Grok app is a separate subscription; the API is billed per token to your xAI account.
• Azure OpenAI16 toolsAzure OpenAI runs OpenAI models inside your own Microsoft Azure subscription. Companies use it for enterprise agreements, data residency and private networking. Apps that support it ask for your Azure endpoint, deployment name and key instead of an OpenAI key.
• Groq16 toolsGroq runs open models such as Llama, Qwen and GPT-OSS on its own LPU hardware and is known for very fast responses. A Groq key is popular in voice and real-time apps where latency matters, and Groq offers a free tier with rate limits.
• LM Studio11 toolsThese tools accept your own LM Studio API key, so usage is billed directly to your LM Studio account instead of through the app's own subscription.
• Amazon Bedrock9 toolsThese tools accept your own Amazon Bedrock API key, so usage is billed directly to your Amazon Bedrock account instead of through the app's own subscription.
• Perplexity7 toolsA Perplexity API key gives access to Sonar models, which search the web and return answers with citations. BYOK apps use it for research and fact-checking features that need current information.
• Meta5 toolsLlama is Meta's family of open-weight models. Apps list Llama when they can run it through a provider key (OpenRouter, Groq, Together AI and others) or locally through Ollama, so you choose where the model runs and what you pay.
• Moonshot AI5 toolsThese tools accept your own Moonshot AI API key, so usage is billed directly to your Moonshot AI account instead of through the app's own subscription.
• Hugging Face4 toolsA Hugging Face token lets apps call models hosted on Hugging Face Inference Providers, including thousands of open models.
• Cerebras3 toolsCerebras serves open models on its wafer-scale chips with very high output speed.
• Cohere3 toolsA Cohere API key gives access to Command models for chat and Embed and Rerank models for search and retrieval, popular in enterprise document search.
• MiniMax3 toolsThese tools accept your own MiniMax API key, so usage is billed directly to your MiniMax account instead of through the app's own subscription.
• Together AI3 toolsA Together AI key gives access to a large catalogue of open models served on Together’s cloud.

Also supported by fewer tools: Fireworks AI (2), Google Vertex AI (2), Qwen (2), Chutes (1), Deepgram (1), ElevenLabs (1), Fal.ai (1), Stability AI (1). Use the search on the home page to find them.