The best AI
tools on earth.
All free. Discover
LLM APIs.
In 2026 you can run production-grade agents, design systems, music studios, and full software teams without spending a single dollar. Here is the living, interactive map.
Why this exists
The free tier revolution is real.
Frontier models, IDEs, voice synthesis, and autonomous agents are now accessible to anyone with an internet connection. This page exists so you never have to pay for AI experimentation again.
The living catalog
Every tool you actually need.
All of them free.
Groq
The fastest inference on the planet. Llama 3.3 70B, Qwen3, Whisper — perfect for real-time agents and live coding.
NotebookLM
Upload PDFs, YouTube, docs → get AI podcasts, study guides, and grounded answers. Magical for research.
n8n
The open-source Zapier killer. Build complex AI workflows, RAG pipelines, and agents that run forever for $0.
Ollama + Continue
Run Qwen3.6-Plus, Llama 4, Gemma locally. Pair with Continue.dev extension for VS Code / JetBrains.
OpenRouter
One key, 300+ models including the newest free gems like MiniMax M2.5 and Nemotron 3 Super.
Cursor
VS Code + GPT-5.1-Codex-Max. The gold standard for AI-native coding. Free tier still incredibly capable.
Ideogram
Still the undisputed king of text-in-images. Magic Prompt turns your rough ideas into stunning posters.
Gemini CLI
Google's official terminal agent with search grounding and MCP. Fast, free, and surprisingly powerful.
ChatGPT
The gateway drug. Sora, DALL·E 4, Deep Research and custom GPTs still free with reasonable limits.
Bing Image Creator
Microsoft's free DALL-E 3 integration. Best quality for zero cost.
Whisper (Local)
OpenAI's speech-to-text. Run entirely offline with excellent accuracy.
Cerebras
Insane 2,400 tokens/sec on Qwen3.6-Plus. The secret weapon for bulk generation and agent swarms.
OpenCode Zen
Curated AI gateway with Big Pickle (72% SWE), MiniMax M2.5 Free, and other top performers.
Qoder
Qwen3.6-Plus (71% SWE) + multi-agent modes. Quest Mode builds entire apps autonomously on free tier.
RooCode
Open-source VS Code extension with a whole team of AI agents. Bring your own API keys.
OpenCode
Go-based open source agent with Big Pickle, MiniMax M2.5 Free and other 72-80% SWE-bench beasts.
MemoryPalace
AI memory system with 96.6% LongMemEval score. Uses memory palace technique for infinite context.
Gemini
Google's frontier chat with native search, 2M context and the best free research agent.
HuggingChat
Open-source chat with the latest community models. No filters, fully transparent.
LMArena
Crowd-sourced arena to compare models. Free access to many LLMs.
Meta AI
Meta's official chat with Llama 4. No limits, integrated with Facebook/Instagram.
Pollinations Chat
Access many models (OpenAI, Claude, Mistral, Gemini) with no signup. Image generation too.
Suno
Write lyrics or hum a melody → complete studio-quality tracks with vocals. Still the best free music AI.
Pollinations Media
Generate images, videos, and music with a simple API. No signup, no watermark.
QuantFlow Pilot
#1 terminal benchmark agent. Paste a GitHub issue → it claims it, plans, codes, tests and opens a PR.
Open WebUI
ChatGPT interface for your local Ollama models. Beautiful, extensible, and completely private.
Continue.dev
The open-source autopilot for VS Code and JetBrains. Bring any local or cloud model you want.
Cline
VS Code extension with Plan/Act modes and MCP support. Bring your own API key or run local models.
LM Studio
Desktop app for running local LLMs. No terminal required, supports GPU acceleration.
Piper
High-quality offline text-to-speech. Fast and lightweight.
FLUX.2
Black Forest Labs' state-of-the-art image model. Excellent prompt adherence.
Playwright
Reliable cross-browser automation. Great for testing and scraping.
Promptfoo
Red-teaming and prompt testing. Catch regressions before deploy.
RAGAS
Evaluate RAG pipelines. Measures faithfulness, answer relevancy, context recall.
Instructor
Pydantic validation for LLM outputs. Makes JSON extraction reliable.
LlamaIndex
Production RAG framework. Smart chunking, reranking, and evaluation.
LangChain
The original LLM orchestration framework. Huge ecosystem of integrations.
BGE-Large-EN
High-quality retrieval embeddings from BAAI. Open source.
Google AI Studio
Gemini 3.1 Flash & Pro with 2M context. Best free multimodal and long-document reasoning.
Cloudflare Workers AI
Edge AI with 40+ models. Run LLMs globally with zero cold starts — perfect for production agents.
Mistral La Plateforme
Access Mistral Large, Small, Codestral. Phone verification required, data training opt-in.
Codestral
Mistral's coding-optimized model. 30 RPM, perfect for code generation and completion.
Trae
VS Code fork with DeepSeek V4 & GPT-4.1. Generous free tier and zero credit card required.
Codeium
The original free powerhouse. 70+ languages, works everywhere. Pro unlocks Claude & GPT-4o.
Void IDE
Agent-first IDE with specialized frontend/backend/testing agents. Describe your app and it builds.
Aider
The original git-native pair programmer. Voice, multi-file edits, and works offline with local models.
Rovo Dev CLI
Atlassian's AI agent with Jira/Confluence integration. Claude Sonnet 4 & GPT-5 preview.
Kilo Code
Open-source VS Code extension with pay-as-you-go, no markup on model pricing.
Claude
Best pure reasoning and writing. The model that actually understands your 100-page codebase.
Phind
Search-first AI for developers. Cites sources, great for technical questions.
Kimi
Moonshot AI's 1M context assistant. Great for long documents, open-source models available.
LobeChat
Aggregator of 40+ models including Gemini, DeepSeek, Claude, GPT. Free tokens monthly.
Pi AI
Inflection's empathetic conversational AI. Great for personal coaching and brainstorming.
G4F
Python library and web UI that scrapes many AI providers. Proceed with caution.
Leonardo.Ai
Commercial-safe images, motion, and game assets. One of the most generous image platforms.
ElevenLabs
The gold standard for voice cloning and TTS. Free tier is enough for most indie projects and demos.
Runway ML
Creative assistant for images, videos, and stories. No signup required for basic use.
Vheer
Proprietary AI image and video generator. Surprisingly good quality.
Julius
Chat with your CSVs and spreadsheets. Instantly generates charts, insights and even forecasts.
Canva AI
Magic Studio: background remover, writer, image gen, video editor — all inside the design tool you already love.
LanguageTool
Open-source grammar checker for 25+ languages. Privacy-focused.
Fathom
Record, transcribe, and summarize Zoom/Teams meetings. Auto-syncs to CRM.
DeepL
High-quality translations with incognito sensitive mode. Best for European languages.
Activepieces
Open-source automation platform with flat pricing. Self-hostable.
ComfyUI
Node-based Stable Diffusion powerhouse. FLUX, SD3, video — the most powerful free image stack on earth.
Deepgram Nova
Superior accuracy for production STT. Handles accents and background noise well.
AssemblyAI
LeMUR for audio RAG, diarization, content moderation.
Bark
Expressive TTS from Suno. Can do singing, laughing, and emotional speech.
Stable Diffusion XL
The original open image model. Huge ecosystem of fine-tunes.
DiffusionBee
Easiest way to run Stable Diffusion on Mac. One-click install.
Steel.dev
Browser API built for LLM agents. Natural language control.
Stagehand
AI browser framework built on Playwright. Natural language commands.
Supabase Vector
Postgres + pgvector. Full-stack vector search with auth and realtime.
LanceDB
Embedded vector DB for multimodal data. Serverless option in beta.
ChromaDB (Cloud)
Simple, Python-native vector DB. Cloud version available.
DeepEval
Unit testing for LLMs. Pytest integration.
Arize Phoenix
LLM observability and evaluation. Traces, evals, and drift detection.
Guidance
Constrained generation with Handlebars syntax. Great for complex schemas.
Outlines
Regex and JSON constrained generation. Fast inference.
Haystack
Enterprise RAG framework. Built for production pipelines.
Vercel AI SDK
Edge-native RAG for Next.js. Streaming, tool calling, and caching.
Jina Embeddings v3
Multilingual embeddings with 8K context. Best for RAG.
E5-Mistral-7B
Best-in-class embedding model. 4096 dimensions, excellent accuracy.
Nomic Embed v1.5
Long context (8K) embeddings. Open-source and fast.
Datawrapper
Publish-ready charts in seconds. Used by journalists.
Looker Studio
Google's BI tool. Seamless with Google Analytics and Ads.
Power BI Desktop
Microsoft's report builder. Copilot recommendations included.
Fotor
AI photo editor and design tool. Background removal, filters, etc.
GitHub Models
Free access to frontier models (including GPT-5 series) right inside your GitHub workflows.
NVIDIA NIM
46 production-grade models including Llama 4 and Mistral Large. Phone verify unlocks serious power.
OfoxAI
Unified API gateway for 100+ LLMs. China-friendly with Hong Kong direct access. OpenAI/Anthropic SDK compatible.
SambaNova Cloud
Fast inference on Llama, Qwen, DeepSeek. Great for low-latency apps.
Perplexity API
Access Sonar models with web search grounding. Great for research agents.
Replicate
Run open-source models in the cloud. Huge catalog of LLMs, images, video.
SiliconFlow
Chinese provider with Qwen, DeepSeek, and more. Great for Asia-based apps.
DeepInfra
Serverless inference for many open models. No credit card required for free tier.
JetBrains AI Assistant
Local models + limited cloud quota. Offline mode with Ollama/LM Studio.
Zed
High-performance editor with native AI. Built in Rust, incredibly fast.
Goose
Block's autonomous desktop + CLI agent. Bring any key and let it ship real engineering work.
Jules
Google's new autonomous coding agent. Turns issues into PRs with Gemini 2.5 Pro.
Warp
AI-powered terminal with GPT-4.1, Claude Opus, Gemini. BYOK option available.
GitHub Copilot CLI
Agent mode with autonomous multi-step coding. Free for VS Code users.
AWS Kiro
AWS-hosted Claude 4 Sonnet. No credit card required for free tier.
Claude Code
Anthropic's terminal agent with extended thinking modes (megathink, ultrathink).
OpenAI Codex CLI
GPT-5.1-Codex-Max with compaction for 24+ hour sessions. Free open-source models available.
Grok
xAI's witty, real-time model with image generation (Aurora) and voice. Surprisingly generous.
DeepSeek Chat
DeepSeek-V3.2 and R1. Excellent reasoning, but politically filtered.
Yupp.ai
Community-driven platform. Earn credits by rating model outputs. Access many models.
Perplexity
AI-powered search with citations. Free tier includes GPT-4o and Claude Sonnet.
Character.AI
Role-playing and character chat. Huge community of user-created bots.
FreeGPT
Community-run proxy offering GPT-4o, DeepSeek, Grok, Gemini.
Recraft
Vector + SVG + brand kits. Perfect for logos, icons, and infinite-resolution design assets.
Krea AI
Real-time image generation and enhancement. Watermark on free tier.
Pixverse
Fast video generation with 540p/750p output. Watermarked.
Grammarly
AI writing assistant with tone detection and rewrites. Great for non-native speakers.
Humata
Chat with PDFs, get answers with clickable citations. Great for research papers.
Make (Integromat)
Beautiful visual builder for AI agents. 1,000 free ops is plenty to automate serious parts of your life.
Gumloop
No-code agent builder. Natural language workflows with 'Gummie' troubleshooting agent.
Relay.app
Simple agentic workflows for beginners. Great for personal automations.
Deepgram Streaming
Low-latency speech-to-text streaming. ~300ms, excellent accuracy.
AssemblyAI Streaming
Realtime speech-to-text with diarization and entity detection.
Sora 3
OpenAI's video model. 120s clips with realistic physics.
Kling
Kuaishou's excellent video model. Supports 2-10 minute videos.
Stable Video Diffusion
Stability's open video model. 4-second clips, fine-tunable.
Browserbase
Managed browsers for AI agents. Session recording, stealth mode.
Qdrant Cloud
High-performance vector database built in Rust. Great for similarity search.
Weaviate Cloud
Hybrid search (vector + keyword). Great for RAG.
LangSmith
Trace, evaluate, and monitor LLM applications. Industry standard.
JSONformer
Structure-aware decoding for local models. Guarantees valid JSON.
OpenAI Text Embedding
Reliable embeddings for general use. Small and large variants.
Modal
Serverless GPU inference. Cold starts in seconds, great for batch jobs.
Vast.ai
Peer-to-peer GPU rental. Very low prices for hobbyists.
Browserbase
Managed browsers for AI agents. Record sessions, replay, and integrate.
Gumloop
No-code agent workflows. Natural language builder.
Relay.app
Simple agentic automations for beginners.
Anomaly AI
Generate interactive dashboards from natural language.
Flourish
No-code interactive maps and 'scrollytelling' visualizations.
AI for Database
Query databases in plain English. Self-refreshing dashboards.
1 More Shot
AI music videos with advanced lip-sync and frame control.
HuggingFace Inference
Serverless inference for models <10GB. Access thousands of open models via a simple API.
Vercel AI Gateway
Route to multiple providers with one API. Built for edge, includes caching and fallbacks.
Cohere
Command R+ 2026, Aya Expanse, and multilingual models. Great for RAG and classification.
Hyperbolic
Access DeepSeek, Llama, Qwen, GPT-OSS with competitive pricing. Simple API.
Fireworks AI
Fast inference on fine-tuned open models. Great for prototyping.
Together AI
19+ open models with generous free credits for new users.
Windsurf
Multi-provider AI IDE with OpenAI, Anthropic, Google, xAI models. Credit card required.
Tabnine
600+ languages, local processing available. Pro adds Claude, GPT-4o.
Lovable
Build web apps with natural language. Free tier good for small projects.
v0.dev
Vercel's AI for UI generation. GPT-5 requires premium.
Amazon Q Developer
AWS's AI coding assistant with Claude Sonnet. Credit card required for free tier.
Poe
Access dozens of models (GPT-4, Claude, Gemini) in one interface. Daily token reset.
Mistral Le Chat
European alternative with structured output and few-shot capabilities.
Adobe Firefly
Commercial-safe image and video generation. Great for Photoshop integration.
Pika
Video generation with lip-sync improvements. Watermarked free tier.
Genmo
Mochi-1 video generation. Quality is decent, watermark included.
Stable Audio
Stability AI's music generation. Good for loops and sound effects.
Riffusion
Generate music from text prompts. Open-source backend.
QuillBot
Paraphrasing and summarization tool. Fluency and standard modes free.
Gemini Live API
Multimodal streaming from Google. Handle audio, video, text in realtime.
Veo 3
Google's cinematic video model. 60s clips at 1080p with sound.
Luma Dream Machine
Fast, realistic video generation. Great for product demos.
Pinecone
Fully managed vector DB. Production-ready, no ops.
RunPod
GPU cloud with spot pricing. Great for training and fine-tuning.
Beam.cloud
Serverless GPU for spiky traffic. Great for demos.
Synthesia
AI video avatars in 120+ languages. Great for training videos.
Bolt.new
Browser-based AI coding with full-stack deployment. Credit card required.
Uberduck
Rapper voice synthesis and text-to-speech with celebrity voices.
Podium
24/7 lead response AI agents for sales and communication.
OpenAI Realtime API
End-to-end realtime voice API. Speech-to-text, LLM, text-to-speech in one stream. ~200ms latency.
Imagen 3
Google's photorealistic image model. Great for product shots.
Lambda Labs
Reliable cloud GPUs for serious work. A100/H100 available.
Midjourney v6
Artistic leader in image generation. No free tier, but included for completeness.
Your personal arsenal
My Free AI Stack (0)
Tap the heart on any tool above to build your personal free stack.
The four story stacks above are a great place to start.
The future is already free
You now have everything you need
to build at the highest level.
The only limit left is your imagination. Go ship something remarkable — and tell us what you made.