AI Tools
Models, agents, and builders that turn prompts into work.
The public hub for models, datasets, and Spaces.
AI video generation and editing in the browser.
Generate full songs from a text prompt.
Replicate
CommunityRun open models with a cloud API
AI ToolsRun and fine-tune models through a simple cloud API.
Generate images and video from a prompt, in a studio.
Generate and edit images, video, and 3D in one studio.
Ask a research question and get papers that actually answer it.
Chat with custom characters, or make one, in a browser tab.
Datasets, notebooks, and ML competitions in one hub.
One API key, hundreds of models. Swap providers without rewriting the client.
Turn a Python script into a shareable data or ML app in a browser tab.
Microsoft’s AI assistant in the browser: write, summarize, and search with the models behind Copilot.
A free chat model that writes, codes, and reasons in the browser.
Generate images, video, and audio from a prompt in Adobe’s Firefly studio.
Run Python notebooks in a tab with a free GPU — no local install.
Generate a full song from a prompt, with stems you can download.
A marketing AI that drafts ads, blogs, and brand copy from a brief.
Google’s Gemini playground: write a prompt, swap models, then grab an API key.
Ask a science question and get answers pulled from peer-reviewed papers.
Describe an app in a prompt and get a running site you can edit in the browser.
Turn a prompt into a video clip, then extend or restyle the shot.
Run open models through a fast inference API — swap Llama, Qwen, and the rest without a GPU rack.
Paraphrase a paragraph, check grammar, and see a fluency rewrite before you send it.
Turn a prompt into a slide deck, doc, or one-page site — then edit the cards.
AI workflows for sales and marketing: research an account, draft the outreach, and generate the page copy.
Ask it, write it, or generate an image — chat, files, and custom GPTs in one tab.
Paste a paper and see a graph of related work — priors, derivatives, and clusters.
Describe an app in chat and get a working web app you can edit and ship.
Enterprise language models you can run in your own cloud — Command, Embed, and Rerank behind an API.
Ask Meta’s assistant in the browser — chat, images, and a voice option on the official Meta AI site.
Search and change code across every repo — then give agents the same full-codebase context.
Run image, video, and audio models through one API without standing up a GPU rack.
OpenAI
The lab behind ChatGPT, the API, and the model cards
AI ToolsThe official desk for ChatGPT, the API, and model cards — start a product or a key from one homepage.
An AI notepad for meetings: record the call, get notes, and skip writing the recap yourself.
Mute the cafe and the other tab. Krisp strips noise from your mic so the call hears you, not the room.
Fathom
Record the call, get the notes — skip writing the recap
AI ToolsRecord the Zoom, Meet, or Teams call and get the notes — a summary and action items instead of typing the recap.
Record Zoom, Meet, or Teams and get searchable notes — a summary and clips instead of typing the recap.
Log training runs, compare models, and trace agent steps — an AI desk instead of a folder of lost notebooks.
Rent a GPU pod or a serverless endpoint — billed while the card actually runs.
Parse messy PDFs into clean text and tables so an agent can actually read the file.
Wafer-scale chips for very fast LLM inference — an OpenAI-compatible API, not a GPU queue.
Run GPU jobs in Python — inference, training, and sandboxes that scale to zero.
Trace LLM calls, prompts, and evals — see what the agent actually did.
An AI gateway: route model traffic, then inspect traces, costs, and failures.
Observability and evals so you can ship quality AI agents instead of guessing from logs.
A vector database for RAG and agents — store embeddings and query nearest neighbors in real time.
AI search for developers — answers with citations, not a chat dump.
Store embeddings and query nearest neighbors for RAG and agents — a hosted vector database.
Open-source vector search for RAG and agents — store embeddings and query them without a custom index stack.
Run open models through a cheap inference API — pay per token, skip standing up a GPU rack.
Run large open models on purpose-built RDUs — fast inference and on-prem stacks without standing up a GPU farm.
Hyperbolic
On-demand H100, H200, and B200 GPUs plus an inference API
AI ToolsRent H100s by the hour or serve a model through an OpenAI-compatible API — GPU capacity in minutes, not a quota ticket.
Scale Ray jobs for training, embeddings, and post-training on your GPUs — a cluster you launch in Python, not a ticket queue.
Free practical deep-learning courses plus the fastai PyTorch library — train a model from a notebook, not a paper recap.
Rent NVIDIA GPU instances or a 1-Click Cluster — training and inference boxes you start in minutes, not a quota ticket.
CoreWeave
An AI cloud of GPU clusters for training and inference
AI ToolsGPU clusters for training and inference on a Kubernetes-native AI cloud — compute, not a quota ticket.
Run open models locally and chat on your machine — a desktop runtime, not a cloud quota.
AI workspace: slides, dashboards, and a Super Agent from a prompt — a desk, not a chat dump.
Route 300+ models through one encrypted inference API — a gateway, not a chat tab.
Chat, code, and agent tasks with Moonshot’s Kimi models — a desk, not a clip dump.
Multimodal models plus Code, Design, Audio, and video tools — a studio, not a chat dump.
Hands-on AI agents that run tasks in a cloud computer — an agent, not a chat dump.
Design and launch conversational AI agents without a custom stack — an agent desk, not a chat dump.
Build AI agents and workflows for work without a custom stack — an agent desk, not a chat dump.
An AI teammate that works in Slack, email, and your apps — an agent desk, not a chat dump.
A no-code desk to build, govern, and deploy enterprise AI agents — a workflow bench, not a chat dump.
Deploy and serve AI models with dedicated inference and model APIs — a runtime, not a chat dump.
Describe a data job and Mage builds, runs, and monitors the workflow — a pipeline desk, not a notebook dump.
A column-store SQL engine for real-time analytics and agent workloads — a database, not a dashboard dump.
Python workflow orchestration for data, ML, and agent pipelines you can observe and retry — a desk, not a notebook dump.
Asset-based data orchestration with lineage, observability, and Dagster+ AI on top of your pipelines — a desk, not a notebook dump.
Snowflake
An AI data cloud for warehouses, pipelines, and agents that query governed enterprise data
AI ToolsAn AI data cloud for warehouses, pipelines, and agents that query governed enterprise data — a database, not a dashboard dump.
A lakehouse for data, analytics, and AI agents on governed enterprise data — a warehouse, not a dashboard dump.
Automated ELT that copies SaaS and database data into your warehouse on a schedule — a pipeline, not a dashboard dump.
Open-source data integration that moves SaaS and database data into warehouses and AI context layers — a pipeline, not a dashboard dump.
Transform warehouse data with SQL models, tests, and docs you version in git — a pipeline, not a dashboard dump.
Google’s app backend: Auth, Firestore, Cloud Functions, and hosting in one project — a pipeline, not a dashboard dump.
Distributed SQL database that survives region failures and scales like a spreadsheet of nodes — a pipeline, not a dashboard dump.
Scale AI
Data labeling and evaluation for teams that train and test production AI systems
AI ToolsData labeling and evaluation for teams that train and test production AI systems — a pipeline, not a dashboard dump.
DataRobot
An enterprise platform to build, govern, and run AI agents on company data
AI ToolsAn enterprise platform to build, govern, and run AI agents on company data — a pipeline, not a dashboard dump.
A data engine for labeling, evaluating, and running RL on the examples that train production models — a pipeline, not a dashboard dump.
Collaborative notebooks and AI analytics that query the warehouse and publish a trusted report — a pipeline, not a dashboard dump.
Track how often AI search cites your brand, then ship the pages and citations that lift it — a pipeline, not a dashboard dump.
Speech-to-text and voice-agent APIs that transcribe audio and pull structured insights from speech — a pipeline, not a dashboard dump.
Speech-to-text, text-to-speech, and a voice-agent API for live conversations, in the cloud or self-hosted — a pipeline, not a dashboard dump.
Ask a research question and see whether later papers support or contradict a finding, with citations you can open — a pipeline, not a dashboard dump.
Build on-brand slide decks with AI outlines and auto-aligning Smart Slides, then share or export to PPT — a desk, not a filter dump.
Claude models, Claude.ai, and an API for writing, coding, and long-context work — a pipeline, not a dashboard dump.
An AI add-on that creates and remixes slides inside PowerPoint and Google Slides — a desk, not a filter dump.
Search, embeddings, and reader APIs that turn the web into tokens models can use — a pipeline, not a dashboard dump.
Embedding and rerank models for retrieval, search, and RAG pipelines — a pipeline, not a dashboard dump.
Cline
Open-source coding agent that edits files, runs commands, and uses tools in your editor
AI ToolsOpen-source coding agent that edits files, runs commands, and uses tools in your editor — a pipeline, not a dashboard dump.
Turn websites into clean markdown and structured data for RAG and agents — a pipeline, not a dashboard dump.
Turn a script into a studio-style voiceover with AI voices and a timeline editor — a pipeline, not a dashboard dump.
Turn a script or prompt into a finished video with AI agents for edit, voice, and b-roll — a pipeline, not a dashboard dump.
Qwen
Chat, code, and generate images with Alibaba Cloud’s open Qwen models from the official studio
AI ToolsChat, code, and generate images with Alibaba Cloud’s open Qwen models from the official studio — a pipeline, not a dashboard dump.
Turn a script, blog, or long video into a short captioned clip with AI edit tools — a pipeline, not a dashboard dump.
Turn a still photo into a talking avatar video for training, support, or a brand presenter — a pipeline, not a dashboard dump.
Give an AI agent a real cloud browser so it can click, type, and finish a web task — a pipeline, not a dashboard dump.
LiteLLM
One OpenAI-compatible gateway to 100+ LLM providers, with spend tracking you can self-host
AI ToolsOne OpenAI-compatible gateway to 100+ LLM providers, with spend tracking you can self-host — a pipeline, not a dashboard dump.
AnythingLLM
A private desktop AI that chats with your files on-device, with no account or API key required
AI ToolsA private desktop AI that chats with your files on-device, with no account or API key required — a pipeline, not a dashboard dump.
GPT4All
Run open-source language models on your own PC, with local document chat and no cloud login required
AI ToolsRun open-source language models on your own PC, with local document chat and no cloud login required — a pipeline, not a dashboard dump.
Run language, vision, and voice models on your own machine behind an OpenAI-compatible API — a pipeline, not a dashboard dump.
Unstructured
Turn PDFs and messy files into clean, chunked data ready for RAG and agents
AI ToolsTurn PDFs and messy files into clean, chunked data ready for RAG and agents — a pipeline, not a dashboard dump.
Chroma
Open-source vector search for embeddings, with a local library and a serverless cloud
AI ToolsOpen-source vector search for embeddings, with a local library and a serverless cloud — a pipeline, not a dashboard dump.
Continue
An open-source coding assistant that lives in your editor and talks to local or cloud models
AI ToolsAn open-source coding assistant that lives in your editor and talks to local or cloud models — a pipeline, not a dashboard dump.
Flowise
Build LLM chatbots and agents on a visual canvas, then self-host or run in the cloud
AI ToolsBuild LLM chatbots and agents on a visual canvas, then self-host or run in the cloud — a pipeline, not a dashboard dump.
Gradio
Build a web UI for a machine-learning model with Python components, then share a link or host it
AI ToolsBuild a web UI for a machine-learning model with Python components, then share a link or host it — a pipeline, not a dashboard dump.
Google DeepMind’s official home for Gemini research, AlphaFold, and the lab’s public models and papers — a pipeline, not a dashboard dump.
Tabby
A self-hosted AI coding assistant that suggests and completes code from your own models or a local box
AI ToolsA self-hosted AI coding assistant that suggests and completes code from your own models or a local box — a pipeline, not a dashboard dump.
Inflection AI
Inflection AI’s official lab site for Pi and the company’s models, research, and product pages
AI ToolsInflection AI’s official lab site for Pi and the company’s models, research, and product pages — a pipeline, not a dashboard dump.
Paperspace
Rent cloud GPUs and notebooks for training or running models, from Paperspace by DigitalOcean
AI ToolsRent cloud GPUs and notebooks for training or running models, from Paperspace by DigitalOcean — a pipeline, not a dashboard dump.
Lambda
Cloud GPUs and AI clusters you can rent by the hour, from a workstation to a training run
AI ToolsCloud GPUs and AI clusters you can rent by the hour, from a workstation to a training run — a pipeline, not a dashboard dump.
Venice
Private AI chat and image generation that runs without keeping a transcript on a company server
AI ToolsPrivate AI chat and image generation that runs without keeping a transcript on a company server — a pipeline, not a dashboard dump.
The official Anaconda platform for Python, packages, and AI notebooks you can run locally or in the cloud — a lab, not a prompt dump.
An open-source platform to track experiments, package models, and ship AI agents from one project log — a lab, not a prompt dump.
AI21
Enterprise language models and AI systems you can build on from the official AI21 site
AI ToolsEnterprise language models and AI systems you can build on from the official AI21 site — a pipeline, not a dashboard dump.