21 tools
Models and platforms
For builders: model APIs, places to run open models locally or in the cloud, and the frameworks and services AI apps are built on.
Start with Hugging Face or Ollama. 15 of 21 can be used for free. Search within platforms
- Hugging FaceFree planThe main hub for open models, datasets and demo apps, with hosted inference and GPUs for running them.
- OllamaOpen sourceAn open-source app and CLI for running open models on your own computer, with optional cloud models.
- CerebrasFree trialAn inference cloud on Cerebras wafer-scale chips with an OpenAI-compatible API for very fast open-model output.
- Claude PlatformPaidAnthropic's API and console for building with Claude models, with managed agents, tool use and prompt caching.
- ExaFree planA web search API built for AI agents, returning results, full page contents and research answers with citations.
- falPaidAn API platform for 1,000+ image, video, audio and 3D generation models, plus serverless GPUs for custom models.
- Google AI StudioFree planGoogle's browser workspace for trying Gemini models, getting an API key and prototyping apps in Build mode.
- GroqFree planAn inference cloud with an OpenAI-compatible API for fast, low-latency responses from open models.
- JanOpen sourceAn open-source desktop chat app that runs AI models offline on your computer or connects to cloud models.
- LangChainOpen sourceOpen-source frameworks for building AI agents, plus LangSmith for tracing, evaluating and deploying them.
- LlamaFreeMeta's family of open-weight language models, from small on-device models to Llama 4, for download and self-hosting.
- llama.cppOpen sourceThe open-source C/C++ engine for running quantized LLMs on almost any hardware, with an OpenAI-compatible server.
- LlamaIndexOpen sourceLlamaParse document parsing and extraction for AI apps, from the makers of the open-source LlamaIndex framework.
- LM StudioFree planA desktop app for downloading and running open models locally, now paired with Bionic, its agent for open models.
- ModalFree planServerless cloud compute that runs Python on GPUs billed by the second, for inference, training and sandboxes.
- OpenAI APIPaidPay-per-token access to OpenAI's GPT models for text, code, images and voice, with APIs for building agents.
- OpenRouterFree planOne OpenAI-compatible API and key for 500+ models from 80+ providers, with automatic fallback between them.
- PineconeFree planA managed, serverless vector database for search and RAG, with hosted embedding and reranking models.
- ReplicateFree trialRun thousands of open and proprietary AI models through an API, paying per second of compute or per output.
- Together AIPaidA cloud for running, fine-tuning and training open models, from serverless APIs to large GPU clusters.
- vLLMOpen sourceAn open-source engine for serving LLMs at high throughput on GPUs, with an OpenAI-compatible API server.