[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-1415":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":13,"stars30d":18,"stars90d":16,"forks30d":16,"starsTrendScore":19,"compositeScore":20,"rankGlobal":10,"rankLanguage":10,"license":21,"archived":22,"fork":22,"defaultBranch":23,"hasWiki":24,"hasPages":22,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":45,"readmeContent":46,"aiSummary":47,"trendingCount":16,"starSnapshotCount":16,"syncStatus":48,"lastSyncTime":49,"discoverSource":50},1415,"free-ai-tools","ShaikhWarsi\u002Ffree-ai-tools","ShaikhWarsi","Curated list of free and low cost AI tools, LLM APIs, IDEs, agents, and infrastructure for building real AI apps","https:\u002F\u002FFreeAiToolsList.vercel.app",null,"TypeScript",429,53,7,3,0,22,180,66,5.2,"MIT License",false,"main",true,[26,27,28,29,30,31,32,33,34,35,36,37,38,5,39,40,41,42,43,44],"ai","ai-agents","ai-directory","ai-tools","ai-tools-directory","ai-tools-list","ai-top-tools","awesome","awesome-ai","awesome-ai-tools","awesome-list","best-ai-tools-2026","developer-resources","free-ai-tools-for-pc","open-source","theresanaiforthat","tools","vibe-coding","workflow","2026-06-12 02:00:27","# Free AI Tools\n\n![Stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FShaikhWarsi\u002Ffree-ai-tools?style=social)\n![Last Updated](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fupdated-April%2011%2C%202026-brightgreen)\n![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-blue)\n![Contributions](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fcontributions-welcome-brightgreen)\n\n> **Curated list of free LLM APIs, coding copilots, AI IDEs, agents, and infrastructure tools for building real AI applications.**\n\n### What's Inside\n- ✅ Free GPT-5 \u002F Claude \u002F Gemini API access\n- 🤖 Coding copilots and AI-native IDEs (Cursor, Trae, Windsurf)\n- 💰 Cheapest AI APIs ($0.10-0.50 per 1M tokens)\n- 📚 RAG stack tools (vector DBs, embeddings, frameworks)\n- 🎯 Agent frameworks and automation tools\n- 🔒 Local models for privacy (Ollama, Llama, Qwen)\n- 🏗️ Production-ready stack configurations\n\n**Goal:** Help developers build AI apps without paying $200\u002Fmonth.\n\n> [!NOTE]  \n> Please don't abuse these services, else we might lose them for everyone.\n\n> [!WARNING]  \n> **April 2026 Model Tier Changes:** Major providers (OpenAI, Anthropic, Google) have restricted flagship models (GPT-5.4, Claude Opus 4.6, Gemini 3.1 Pro) to paid tiers. Free tiers now get lighter models (GPT-4o, Claude Sonnet\u002FHaiku, Gemini Flash). Entries marked with `[verify]` need confirmation.\n\n---\n\n## 🎯 Why This Repo Exists\n\nMost AI tool lists are:\n- ❌ Outdated (prices\u002Flimits from 2023)\n- ❌ Filled with affiliate links and sponsored placements\n- ❌ General-purpose directories with no developer focus\n- ❌ Missing production-critical details (rate limits, commercial use, architecture patterns)\n\n**This repo focuses only on:**\n- ✅ Tools developers *actually* use in production\n- ✅ Generous free tiers (no \"5 requests then paywall\")\n- ✅ Production-capable models (SWE-bench verified, not toys)\n- ✅ Real infrastructure (APIs, hosting, vector DBs, not just chatbots)\n- ✅ Minimal fluff, maximum utility\n\n**Unlike:** `awesome-ai` (general list), `ai-collection` (marketing focus), `toolify` (affiliate-heavy)\n\n**This is for:** Builders who want to ship AI features this week.\n\n---\n\n## ⭐ Support This Project\n\nIf this repo helped you build something or saved you money:\n\n**[⭐ Star this repo](https:\u002F\u002Fgithub.com\u002FShaikhWarsi\u002Ffree-ai-tools)** — it helps more builders discover free AI resources.\n\n**[🔄 Share with your team]** — spread the knowledge.\n\n**[📝 Contribute](CONTRIBUTING.md)** — found a new free tier? Updated pricing? PRs welcome!\n\n---\n\n## 📅 Updates\n\n**2026-04-12**\n- ✨ added a website for easy navigation\n---\n**2026-04-11**\n- ✨ Initial release\n---\n\n## Table of Contents\n\n- [Quick Comparison](#quick-comparison)\n- [Free LLM API Providers](#free-llm-api-providers)\n  - [Fully Free Providers](#fully-free-providers)\n  - [Providers with Trial Credits](#providers-with-trial-credits)\n- [AI-Powered IDEs](#ai-powered-ides)\n  - [IDEs with Pro-Grade Models](#ides-with-pro-grade-models)\n  - [IDEs with Basic Models](#ides-with-basic-models)\n- [CLI Coding Tools](#cli-coding-tools)\n  - [CLI Tools with Pro-Grade Models](#cli-tools-with-pro-grade-models)\n  - [CLI Tools with Basic Models](#cli-tools-with-basic-models)\n- [API Providers for AI Coding Tools](#api-providers-for-ai-coding-tools)\n- [Paid Tiers Comparison](#paid-tiers-comparison)\n- [Local Models](#local-models)\n- [free-coding-models CLI](#free-coding-models-cli)\n- [Additional 2026 AI Tools](#additional-2026-ai-tools)\n  - [Agentic Workflow Platforms](#agentic-workflow-platforms)\n  - [Data Visualization & Analysis](#data-visualization--analysis)\n  - [Creative & Multimedia Tools](#creative--multimedia-tools)\n  - [Productivity & Research Tools](#productivity--research-tools)\n  - [Vertical AI](#vertical-ai)\n  - [Marketing & SEO Tools](#marketing--seo-tools)\n  - [Open Source & Local Tools](#open-source--local-tools)\n- [🏗️ Recommended Stacks](#️-recommended-stacks)\n- [⚡ Realtime & Streaming APIs](#-realtime--streaming-apis)\n- [🎙️ Speech Models](#️-speech-models)\n- [🎨 Image Generation Models](#-image-generation-models)\n- [🎬 Video Generation APIs](#-video-generation-apis)\n- [🌐 AI Browser Automation](#-ai-browser-automation)\n- [💾 Cheap Vector DB Hosting](#-cheap-vector-db-hosting)\n- [🏛️ Common AI Architecture Patterns](#️-common-ai-architecture-patterns)\n- [💵 Model Price Comparison](#-model-price-comparison)\n- [🎯 Best Models by Use Case](#-best-models-by-use-case)\n- [⏱️ Rate Limit Comparison](#️-rate-limit-comparison)\n- [✅ Commercial Use Summary](#-commercial-use-summary)\n- [🧩 RAG Stack Tools](#-rag-stack-tools)\n- [🔢 Best Free Embedding APIs](#-best-free-embedding-apis)\n- [🖥️ AI Hosting & GPU Providers](#️-ai-hosting--gpu-providers)\n- [📊 AI Evaluation Tools](#-ai-evaluation-tools)\n- [📐 Structured Output Tools](#-structured-output-tools)\n- [🏷️ Legend](#️-legend)\n- [Contributing](#contributing)\n- [License](#license)\n\n---\n\n## Quick Comparison\n\n### Free LLM API Providers Summary\n\n| Provider | Models | Free Tier | Credit Card |\n|----------|--------|-----------|-------------|\n| [NVIDIA NIM](#nvidia-nim) | 46 | 40 req\u002Fmin | No |\n| [OpenRouter](#openrouter) | 25 | 50\u002Fday (1K\u002Fday with $10) | No |\n| [Groq](#groq) | 20+ | 1K-14.4K req\u002Fday | No |\n| [Google AI Studio](#google-ai-studio) | 9 | 5-500 req\u002Fday | No |\n| [Cloudflare Workers AI](#cloudflare-workers-ai) | 47+ | 10K neurons\u002Fday | No |\n| [Cerebras](#cerebras) | 4 | 1M tokens\u002Fday | No |\n| [Cohere](#cohere) | 14 | 1K req\u002Fmonth | No |\n| [Mistral La Plateforme](#mistral-la-plateforme) | 10+ | 1B tokens\u002Fmonth | No |\n| [GitHub Models](#github-models) | 30+ | 50 chat + 2K completions\u002Fmonth | No |\n| [SambaNova](#sambanova-cloud) | 13 | $5 for 3 months | No |\n| [Hyperbolic](#hyperbolic) | 13 | $1 trial | No |\n\n### AI-Powered IDEs with Free Pro-Grade Access\n\n| IDE | Pro-grade Models | Free Tier Limit | Credit Card |\n|-----|------------------|-----------------|-------------|\n| [Cursor](#cursor) | GPT-5.1-Codex-Max | Limited free tier | No |\n| [Trae](#trae) | DeepSeek V4, GPT-4.1 (Claude removed Nov 2025) | 10 fast + 50 slow\u002Fmonth | No |\n| [Windsurf](#windsurf) | OpenAI, Anthropic, Google, xAI | 25 credits\u002Fmonth | Required |\n| [Qoder](#qoder) | Qwen3.6-Plus, Qwen3-Coder-480B, Claude, GPT, Gemini | Unlimited completions + limited chat | No |\n\n### CLI Coding Tools with Free Pro-Grade Access\n\n| Tool | Pro-grade Models | Free Tier Limit | Credit Card |\n|------|------------------|-----------------|-------------|\n| [Gemini CLI](#gemini-cli) | Gemini 3.1 Flash [verify: Pro paid] | 100-250 req\u002Fday | No |\n| [Rovo Dev CLI](#rovo-dev-cli) | Claude Sonnet 4 [verify], GPT-5 preview [verify] | 5M tokens\u002Fday | No |\n| [Warp](#warp) | GPT-4.1, Claude Opus 4.1 [verify] | 150 credits\u002Fmonth | No |\n| [GitHub Copilot](#github-copilot) | GPT-4.1, Claude Opus | 50 chat + 2K completions\u002Fmonth | No |\n| [Jules](#jules) | Gemini 2.5 Pro | 15 tasks\u002Fday | No |\n| [AWS Kiro](#aws-kiro) | Claude Sonnet 4 [verify] | 50 credits\u002Fmonth | No |\n| [OpenCode](#opencode) | 300+ models via OpenRouter | Zen Free tier | No |\n| [ForgeCode](#forgecode) | 300+ models via OpenRouter | 10K tokens\u002Fday | No |\n| [Amazon Q Developer](#amazon-q-developer) | Claude Sonnet 4 [verify] | 50 agentic req\u002Fmonth | Required |\n| [RooCode](#roocode) | Bring your own keys | Unlimited (BYOK) | No |\n| [Goose](#goose) | Bring your own keys | Unlimited (BYOK) | No |\n| [OhMyPi](#ohmypi) | Bring your own keys | Unlimited (BYOK) | No |\n\n### What Qualifies as \"Pro-Grade\"?\n\nModels achieving ≥60% on SWE-bench Verified:\n\n| Model | SWE-bench | Provider |\n|-------|-----------|----------|\n| Claude **Opus 4.6** | 84.2% | Anthropic |\n| **GPT-5.4** | 80.1% | OpenAI |\n| Claude Sonnet 4.6 | 79.3% | Anthropic |\n| Gemini 3.1 Pro | 77.4% | Google |\n| Claude Opus 4.5 | 82.1% | Anthropic |\n| GPT-5.1-Codex-Max | 78.3% | OpenAI |\n| Qwen3.6-Plus | 71.2% | Alibaba |\n| Claude Sonnet 4.5 | 77.8% | Anthropic |\n\n> **Note:** `[verify]` indicates scores need verification from official sources. Always check current benchmarks before making decisions.\n\n---\n\n## 🏗️ Recommended Stacks\n\nReady-made combinations for different use cases. Copy-paste these configurations.\n\n### 🟢 Fully Free Coding Stack (No Credit Card)\n\n| Layer | Tool | Why |\n|-------|------|-----|\n| **IDE** | Cursor Hobby \u002F Qoder | GPT-5.4 limited credits |\n| **CLI** | Gemini CLI (3.1 Pro) \u002F Rovo | 100-250 req\u002Fday, 5M tokens\u002Fday |\n| **API** | OpenRouter + Groq | 50 req\u002Fday + 14.4K req\u002Fday combo |\n| **Local** | Ollama + Qwen3.6-Plus | Unlimited offline |\n| **Automation** | n8n Self-hosted | Unlimited workflows |\n| **Vector DB** | ChromaDB \u002F LanceDB | Free local storage |\n\n**Total Cost: $0\u002Fmonth**\n\n---\n\n### ⚡ Fastest Stack (Low Latency)\n\n| Layer | Tool | Speed |\n|-------|------|-------|\n| **Inference** | Groq \u002F Cerebras | 2,000 tokens\u002Fsec (Cerebras) |\n| **Coding** | Qwen3.6-Plus via Groq | 1,000 req\u002Fday (71.2% SWE) |\n| **Agent** | OpenCode Zen | Big Pickle (72.0%), MiniMax M2.5 (80.2%) |\n| **Cache** | DeepSeek V4 | $0.30\u002F$0.50 per 1M, 90% cache discount |\n| **Edge** | Cloudflare Workers AI | Global CDN |\n\n**Best for:** Real-time apps, trading bots, live coding assistants\n\n---\n\n### 💰 Cheapest Pro Stack (\u003C$10\u002Fmonth)\n\n| Layer | Tool | Cost |\n|-------|------|------|\n| **IDE** | Trae Pro | $10\u002Fmo (600 fast, DeepSeek V4\u002FGPT-5.4) |\n| **API** | OpenRouter $10 | 1K req\u002Fday + BYOK 1M\u002Fmonth free |\n| **CLI** | Gemini CLI | v0.37.1 (Gemini 3.1 Pro\u002FFlash) |\n| **Local** | Ollama | Free |\n| **Embeddings** | Jina AI | Free tier |\n\n**Total Cost: ~$10\u002Fmonth for pro-grade everything**\n\n---\n\n### 🔒 Local Privacy Stack (100% Offline)\n\n| Layer | Tool | Privacy |\n|-------|------|---------|\n| **Models** | Ollama + Llama 3.3 \u002F Qwen3-Coder | Runs locally |\n| **IDE** | Continue.dev + VS Code | BYO local models |\n| **CLI** | Aider + local Ollama | Git-integrated, offline |\n| **Chat UI** | Open WebUI | Self-hosted ChatGPT alternative |\n| **Vector DB** | ChromaDB \u002F LanceDB | Local embeddings storage |\n| **Speech** | Whisper (local) | Offline transcription |\n\n**Best for:** Healthcare, legal, finance - any sensitive data\n\n---\n\n### 🤖 Agentic AI Stack (Autonomous Workflows)\n\n| Component | Tool | Role |\n|-----------|------|------|\n| **Orchestrator** | n8n \u002F Gumloop | Workflow automation |\n| **Reasoning** | DeepSeek R1 \u002F DeepSeek V4 | Complex decision making |\n| **Execution** | Qwen3.6-Plus | Code generation |\n| **Memory** | ChromaDB \u002F Supabase Vector | Long-term context |\n| **Embeddings** | Jina Embeddings v3 (1M tokens\u002Fday free) | Semantic search |\n| **Monitoring** | LangSmith | Trace agent steps |\n\n**Best for:** Autonomous research assistants, code review bots, data processing pipelines\n\n---\n\n### 📊 RAG Stack (Document Q&A)\n\n| Component | Tool | Purpose |\n|-----------|------|---------|\n| **Framework** | LlamaIndex \u002F LangChain | RAG orchestration |\n| **Vector DB** | ChromaDB \u002F Weaviate \u002F Supabase | Document storage |\n| **Embeddings** | E5-Mistral-7B (best accuracy) | Text vectorization |\n| **Chunking** | LlamaIndex | Smart document splitting |\n| **Reranking** | Cohere Rerank | Improve retrieval accuracy |\n| **LLM** | Claude Sonnet 4.6 (79.3%) \u002F GPT-5.4 | Answer generation |\n| **Eval** | RAGAS | Measure RAG performance |\n\n**Best for:** ExamAi, legal document analysis, knowledge bases\n\n---\n\n## Free LLM API Providers\n\n### Fully Free Providers\n\n#### [OpenRouter](https:\u002F\u002Fopenrouter.ai)\n\n**Limits:** 20 RPM, **29 free models** (262K context max, March 2026), models share quota\n\n- [Llama 3.3 70B](https:\u002F\u002Fopenrouter.ai\u002Fmeta-llama\u002Fllama-3.3-70b-instruct:free) ✅\n- **NEW: [Nemotron 3 Super](https:\u002F\u002Fopenrouter.ai\u002Fnvidia\u002Fnemotron-3-super:free)** (262K context)\n- **NEW: [MiniMax M2.5](https:\u002F\u002Fopenrouter.ai\u002Fminimax\u002Fminimax-m2.5:free)**\n- **NEW: [Devstral 2](https:\u002F\u002Fopenrouter.ai\u002Fmistralai\u002Fdevstral-2:free)** (Apache 2.0)\n- **NEW: [Gemma 3n family](https:\u002F\u002Fopenrouter.ai\u002Fgoogle\u002Fgemma-3n-e2b-it:free)** (mobile-optimized)\n- **qwen\u002Fqwen3.6-plus:free** ✅\n- [Hermes 3 Llama 3.1 405B](https:\u002F\u002Fopenrouter.ai\u002Fnousresearch\u002Fhermes-3-llama-3.1-405b:free)\n- [Llama 3.2 3B Instruct](https:\u002F\u002Fopenrouter.ai\u002Fmeta-llama\u002Fllama-3.2-3b-instruct:free)\n- [Mistral Small 3.1 24B](https:\u002F\u002Fopenrouter.ai\u002Fmistralai\u002Fmistral-small-3.1-24b-instruct:free)\n- [Full list](https:\u002F\u002Fopenrouter.ai\u002Fcollections\u002Ffree-models)\n\n---\n\n#### [Google AI Studio](https:\u002F\u002Faistudio.google.com)\n\nData is used for training when used outside UK\u002FCH\u002FEEA\u002FEU.\n\n**Rate limits:** Tier 1 (default): 250 RPD | Tier 2: Requires $250 spend + 30 days\n\n| Model | Free Tier Limits |\n|-------|------------------|\n| Gemini 3.1 Pro [verify: now paid] | 250 RPD (Tier 1) |\n| Gemini 3 Flash | 1,500 RPD |\n| All others | Check console |\n\n> **Note:** Data training outside UK\u002FCH\u002FEEA\u002FEU still applies.\n\n---\n\n#### [NVIDIA NIM](https:\u002F\u002Fbuild.nvidia.com\u002Fexplore\u002Fdiscover)\n\nPhone number verification required. Models tend to be context window limited.\n\n**Limits:** **1K credits signup, up to 5K total, 40 RPM** (phone verify required)\n\n- 46+ models including Llama 3.3 70B, Llama 4 Scout, Mistral Large, Qwen3 235B\n\n---\n\n#### [Mistral (La Plateforme)](https:\u002F\u002Fconsole.mistral.ai\u002F)\n\n*Free tier requires opting into data training; phone verification required*\n\n**Limits (per-model):** 1 req\u002Fs, 500K tokens\u002Fmin, 1B tokens\u002Fmonth\n\n- Open and Proprietary Mistral models (Mistral Large 3, Small 3.1, etc.)\n\n---\n\n#### [Mistral (Codestral)](https:\u002F\u002Fcodestral.mistral.ai\u002F)\n\n**Limits:** 30 RPM, 2K RPD confirmed free\n\n- Codestral (monthly subscription-based, currently free)\n\n---\n\n#### [HuggingFace Inference Providers](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Finference-providers\u002Fen\u002Findex)\n\nServerless Inference limited to models \u003C10GB (some popular models >10GB supported).\n\n**Limits:** ~$0.10\u002Fmonth in credits\n\n- Various open models across supported providers\n\n---\n\n#### [Vercel AI Gateway](https:\u002F\u002Fvercel.com\u002Fdocs\u002Fai-gateway)\n\nRoutes to various supported providers.\n\n**Limits:** $5\u002Fmonth\n\n---\n\n#### [OpenCode Zen](https:\u002F\u002Fopencode.ai\u002Fdocs\u002Fzen\u002F)\n\nAI gateway with curated models. Free models may use data for improvement.\n\n- Big Pickle Stealth (S+, 72.0% SWE-bench)\n- MiniMax M2.5 Free (S+, 80.2% SWE-bench)\n- MiMo V2 Pro\u002FOmni\u002FFlash Free\n- Nemotron 3 Super Free\n- GPT 5 Nano\n- Trinity Large Preview Free\n\n---\n\n#### [Cerebras](https:\u002F\u002Fcloud.cerebras.ai\u002F)\n\n| Model | Limits |\n|-------|--------|\n| GPT-OSS 120B | 30 req\u002Fmin, 60K tokens\u002Fmin, 900 req\u002Fhour, 1M tokens\u002Fday |\n| Llama 3.1 8B | Same limits as above |\n| Qwen3-235B | Available via API |\n\n---\n\n#### [Groq](https:\u002F\u002Fconsole.groq.com)\n\n| Model | Limits |\n|-------|--------|\n| Llama 3.1 8B | 14,400 req\u002Fday, 6K tokens\u002Fmin |\n| Llama 3.3 70B | 1,000 req\u002Fday, 12K tokens\u002Fmin |\n| Llama 4 Maverick\u002FScout | 1,000 req\u002Fday |\n| Whisper Large v3\u002Fv3 Turbo | 7,200 audio-sec\u002Fmin, 2,000 req\u002Fday |\n| Qwen3-32B | 1,000 req\u002Fday, 6K tokens\u002Fmin |\n| Kimi K2 Instruct | 1,000 req\u002Fday, 10K tokens\u002Fmin |\n| GPT-OSS 20B\u002F120B | 1,000 req\u002Fday, 8K tokens\u002Fmin |\n| And 15+ more |\n\n---\n\n#### [Cohere](https:\u002F\u002Fcohere.com)\n\n**Limits:** **20 RPM, 1K req\u002Fmonth (non-commercial only)**\n\n- Command R+ 2026\n- c4ai-aya-expanse\u002Fvision-32b\n- command-a\u002Fr\u002Fr7b variants\n\n---\n\n#### [GitHub Models](https:\u002F\u002Fgithub.com\u002Fmarketplace\u002Fmodels)\n\nExtremely restrictive input\u002Foutput token limits.\n\n**Limits:** [Dependent on Copilot subscription tier (Free\u002FPro\u002FPro+\u002FBusiness\u002FEnterprise)](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fgithub-models\u002Fprototyping-with-ai-models#rate-limits)\n\n- AI21 Jamba 1.5 Large\n- Codestral 25.01\n- Cohere Command A, Command R\u002FR+ 08-2024\n- DeepSeek-R1, DeepSeek-R1-0528, DeepSeek-V3.2, DeepSeek-V3-0324\n- Grok 3, Grok 3 Mini\n- Llama 4 Maverick 17B 128E Instruct FP8, Llama 4 Scout 17B 16E Instruct\n- Llama-3.2-11B\u002F90B-Vision-Instruct, Llama-3.3-70B-Instruct\n- MAI-DS-R1, Meta-Llama-3.1-405B\u002F8B-Instruct\n- Ministral 3B, Mistral Medium 3 (25.05), Mistral Small 3.1\n- OpenAI GPT-4.1\u002Fmini\u002Fnano, GPT-4o\u002Fmini, GPT-5\u002Fmini\u002Fnano\n- OpenAI gpt-5-chat (preview), o1\u002Fo1-mini\u002Fo1-preview, o3\u002Fo3-mini, o4-mini\n- OpenAI Text Embedding 3 (large\u002Fsmall)\n- Phi-4, Phi-4-mini-instruct\u002Freasoning, Phi-4-multimodal-instruct, Phi-4-reasoning\n\n---\n\n#### [Cloudflare Workers AI](https:\u002F\u002Fdevelopers.cloudflare.com\u002Fworkers-ai)\n\n**Limits:** [10,000 neurons\u002Fday](https:\u002F\u002Fdevelopers.cloudflare.com\u002Fworkers-ai\u002Fplatform\u002Fpricing\u002F#free-allocation)\n\n- @cf\u002Faisingapore\u002Fgemma-sea-lion-v4-27b-it\n- @cf\u002Fibm-granite\u002Fgranite-4.0-h-micro\n- @cf\u002Fopenai\u002Fgpt-oss-120b, @cf\u002Fopenai\u002Fgpt-oss-20b\n- @cf\u002Fqwen\u002Fqwen3-30b-a3b-fp8\n- @cf\u002Fzai-org\u002Fglm-4.7-flash\n- DeepSeek R1 Distill Qwen 32B\n- Deepseek Coder 6.7B Base\u002FInstruct (AWQ)\n- Deepseek Math 7B Instruct\n- Gemma 2B\u002F3 12B\u002F7B Instruct (LoRA)\n- Hermes 2 Pro Mistral 7B\n- Llama 2 7B\u002F13B Chat (FP16\u002FINT8\u002FAWQ\u002FLoRA)\n- Llama 3 8B Instruct, Llama 3.1 8B Instruct (AWQ\u002FFP8)\n- Llama 3.2 1B\u002F3B\u002F11B Vision Instruct\n- Llama 3.3 70B Instruct (FP8), Llama 4 Scout Instruct\n- Mistral 7B Instruct v0.1\u002Fv0.2 (AWQ\u002FLoRA)\n- Mistral Small 3.1 24B Instruct\n- Qwen 1.5 0.5B\u002F1.8B\u002F7B\u002F14B Chat (AWQ)\n- Qwen 2.5 Coder 32B Instruct, Qwen QwQ 32B\n- Phi-2, SQLCoder 7B 2\n- And more...\n\n---\n\n### Providers with Trial Credits\n\n| Provider | Credits | Duration | Notes |\n|----------|---------|----------|-------|\n| [Fireworks](https:\u002F\u002Ffireworks.ai\u002F) | $1 | Permanent | Various open models |\n| [Baseten](https:\u002F\u002Fapp.baseten.co\u002F) | $30 | Permanent | Pay by compute time |\n| [Nebius](https:\u002F\u002Ftokenfactory.nebius.com\u002F) | $1 | Permanent | Various open models |\n| [Novita](https:\u002F\u002Fnovita.ai\u002F) | $0.50 | 1 year | Various open models |\n| [AI21](https:\u002F\u002Fstudio.ai21.com\u002F) | $10 | 3 months | Jamba family |\n| [Upstage](https:\u002F\u002Fconsole.upstage.ai\u002F) | $10 | 3 months | Solar Pro\u002FMini |\n| [NLP Cloud](https:\u002F\u002Fnlpcloud.com\u002Fhome) | $15 | Permanent | Phone verification required |\n| [Alibaba Cloud](https:\u002F\u002Fbailian.console.alibabacloud.com\u002F) | 1M tokens\u002Fmodel | 90 days | Qwen models |\n| [Modal](https:\u002F\u002Fmodal.com) | $5-30\u002Fmonth | Monthly | Pay by compute time |\n| [Inference.net](https:\u002F\u002Finference.net) | $1 (+$25 on survey) | Permanent | Various open models |\n| [Hyperbolic](https:\u002F\u002Fapp.hyperbolic.ai\u002F) | $1 | Permanent | DeepSeek, Llama, Qwen, GPT-OSS |\n| [SambaNova Cloud](https:\u002F\u002Fcloud.sambanova.ai\u002F) | $5 | 3 months | Llama, Qwen, DeepSeek |\n| [Scaleway](https:\u002F\u002Fconsole.scaleway.com\u002Fgenerative-api\u002Fmodels) | 1M tokens | Permanent | DeepSeek, Llama, Mistral, Gemma |\n\n### Additional Free API Providers\n\n| Provider | Models | Free Tier | Environment Variable |\n|----------|--------|-----------|---------------------|\n| [Together AI](https:\u002F\u002Fapi.together.ai\u002Fsettings\u002Fapi-keys) | 19 | Credits\u002Fpromos vary by account | `TOGETHER_API_KEY` |\n| [iFlow](https:\u002F\u002Fplatform.iflow.cn) | 11 | Free for individuals (7-day key expiry) | `IFLOW_API_KEY` |\n| [ZAI](https:\u002F\u002Fz.ai) | 7 | Free tier (generous quota) | `ZAI_API_KEY` |\n| [SiliconFlow](https:\u002F\u002Fcloud.siliconflow.cn\u002Faccount\u002Fak) | 6 | 1K RPM, 50K TPM | `SILICONFLOW_API_KEY` |\n| [Perplexity API](https:\u002F\u002Fwww.perplexity.ai\u002Fsettings\u002Fapi) | 4 | ~50 RPM default | `PERPLEXITY_API_KEY` |\n| [OVHcloud AI Endpoints](https:\u002F\u002Fendpoints.ai.cloud.ovh.net) | 8 | 2 req\u002Fmin (no key), 400 RPM with key | `OVH_AI_ENDPOINTS_ACCESS_TOKEN` |\n| [Chutes AI](https:\u002F\u002Fchutes.ai) | 4 | Free community GPU-powered | `CHUTES_API_KEY` |\n| [DeepInfra](https:\u002F\u002Fdeepinfra.com\u002Flogin) | 4 | 200 concurrent requests | `DEEPINFRA_API_KEY` |\n| [Replicate](https:\u002F\u002Freplicate.com\u002Faccount\u002Fapi-tokens) | 2 | 6 req\u002Fmin (no payment), up to 3K RPM with payment | `REPLICATE_API_TOKEN` |\n\n---\n\n## AI-Powered IDEs\n\nFull-featured integrated development environments with built-in AI assistance.\n\n### IDEs with Pro-Grade Models\n\n#### [Cursor](https:\u002F\u002Fcursor.com\u002F)\n\n**Model:** GPT-5.1-Codex-Max (77.9% SWE-bench Verified) [verify]\n- **Free tier:** **500 slow premium req\u002Fmo, 2K completions\u002Fmo** (post-Dec 2025 credits)\n- Free models: Cursor Small, Deepseek v3, Gemini 2.5 Flash, GPT-4o mini (500\u002Fday limit), Grok 3 Mini Beta [verify: GPT-5.4 now paid-only]\n- Claude models removed from free tier ~June 2025\n- Free tier uses token-based usage tracking (not request-based)\n- AI-powered code editor with autonomous coding capabilities\n- **Pro ($20\u002Fmo or $16\u002Fmo annually):** Extended Agent limits + Unlimited Tab completions + Background Agents + Maximum context windows\n- **Pro+ ($60\u002Fmo):** 3x usage on all OpenAI, Claude, Gemini models\n- **Ultra ($200\u002Fmo):** 20x usage on all models + Priority access to new features\n- **Teams ($40\u002Fuser\u002Fmo):** Pro features + Centralized billing + Usage analytics + SAML\u002FOIDC SSO\n- **Enterprise (Custom):** Everything in Teams + Pooled usage + SCIM + AI code tracking API + Audit logs\n\n**[Pricing](https:\u002F\u002Fcursor.com\u002Fen\u002Fpricing)** | **[GPT-5.1-Codex-Max Announcement](https:\u002F\u002Fforum.cursor.com\u002Ft\u002Fgpt-5-1-codex-max-available-in-cursor\u002F145277)**\n\n---\n\n#### [Trae](https:\u002F\u002Ftrae.ai\u002F)\n\n**Models:** DeepSeek V4, GPT-4.1, GPT-4o, Gemini 2.5 Pro (Claude models removed Nov 2025)\n- 10 fast requests + 50 slow requests\u002Fmonth for premium models\n- 1,000 slow requests\u002Fmonth for advanced models\n- 5,000 auto-completions\u002Fmonth\n- VS Code-based IDE with AI integration\n- No credit card required for free tier\n- **Pro ($10\u002Fmo):** 600 fast + unlimited slow requests for premium models\n- Unlimited slow requests for advanced models\n- Zero rate limits and faster access to premium models\n- Extra packages available: $3-$12 for additional fast requests\n- First month available for $3\n\n**[Pricing](https:\u002F\u002Ftrae.ai\u002Fpricing)** | **[Documentation](https:\u002F\u002Fdocs.trae.ai\u002Fide\u002Fbilling)**\n\n---\n\n#### [Windsurf](https:\u002F\u002Fwindsurf.com\u002F)\n\n**Models:** OpenAI, Anthropic, Google, xAI model access\n- 25 prompt credits\u002Fmonth limit\n- Multiple providers (OpenAI, Claude, Gemini, xAI)\n- Credit card required\n- Can purchase add-on credits to continue\n- **Pro ($15\u002Fmo):** 500 prompt credits\u002Fmonth\n- **Teams ($30\u002Fuser\u002Fmo):** 500 prompt credits\u002Fuser\u002Fmonth\n- **Enterprise ($60+\u002Fuser\u002Fmo):** 1,000 prompt credits\u002Fuser\u002Fmonth\n\n**[Pricing](https:\u002F\u002Fwindsurf.com\u002Fpricing)**\n\n---\n\n#### [Void IDE](https:\u002F\u002Fvoideditor.com\u002F)\n\n**Models:** Multi-agent (frontend\u002Fbackend\u002Ftesting agents)\n- **Agent-first IDE** - new 2026 category\n- Multiple specialized agents coordinate across codebase\n- Free preview tier with high usage limits\n- VS Code-based\n\n**Best for:** Full-stack development with natural language direction\n\n---\n\n#### [Qoder](https:\u002F\u002Fqoder.com\u002F)\n\n**Models:** Qwen3.6-Plus (71.2% SWE), Qwen-Coder-Qoder, GPT-4o, Claude Sonnet [verify: flagship models now paid-only]\n- **Free tier:** Unlimited completions + **limited chat\u002Fagent (basic models)** + **2-week Pro trial (1,000 credits)**\n- **Experts Mode:** Multi-agent collaboration (new Mar 2026)\n- **Quest Mode:** Fully autonomous app building\n- **Nextnew:** Tab predictions\n- Windows\u002FmacOS, VS Code-based\n\n**Pricing (50% discount - Apr 2026):**\n- **Free:** Basic models, limited messages\n- **Pro:** $10\u002Fmo (reg $20) - **2,000 credits**\n- **Pro+:** $30\u002Fmo (reg $60) - **6,000 credits**\n- **Ultra:** $100\u002Fmo (reg $200)\n- **Credits:** $0.01\u002Fcredit (reg $0.02), expire 1mo\n\n**[Docs](https:\u002F\u002Fdocs.qoder.com\u002F)** | **[Pricing](https:\u002F\u002Fqoder.com\u002Fpricing)**\n\n---\n\n#### [RooCode](https:\u002F\u002Fgithub.com\u002FRooCodeInc\u002FRoo-Code)\n\n**Models:** Bring your own API keys (any provider)\n- Open-source AI-powered coding assistant for VS Code\n- Whole dev team of AI agents in your editor\n- No subscription required - pay-as-you-go with your own keys\n- Custom modes for different coding tasks\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002FRooCodeInc\u002FRoo-Code)** | **[Website](https:\u002F\u002Froocode.com)**\n\n---\n\n### IDEs with Basic Models\n\n#### [Codeium](https:\u002F\u002Fcodeium.com\u002F)\n\n**Model:** Base model (Llama 3.1 70B), pro-grade models require subscription\n- Individual plan: Free forever with unlimited code completions, AI chat, commands\n- 70+ programming languages supported\n- IDE integrations: VS Code, JetBrains, Vim\u002FNeovim, Jupyter\n- No credit card required\n- Limited context awareness (expanded in paid tiers)\n- **Pro ($10\u002Fmo):** Unlimited usage with advanced context awareness, Claude 3.5 Sonnet, GPT-4o access\n- **Teams ($12\u002Fuser\u002Fmo):** Pro features + team management\n- **Enterprise (Custom):** On-premise deployment, custom models\n\n**[Pricing](https:\u002F\u002Fcodeium.com\u002Fpricing)** | **[Documentation](https:\u002F\u002Fcodeium.com\u002Fdocs)**\n\n---\n\n#### [JetBrains AI Assistant](https:\u002F\u002Fwww.jetbrains.com\u002Fai\u002F)\n\n**Models:** Local models + cloud models with limited quota\n- AI Free tier included with IDEs\n- Unlimited code completion and local model support\n- Limited quota for cloud-based features\n- 30-day AI Pro trial included\n- Offline mode with local models via Ollama\u002FLM Studio\n- **AI Pro ($15\u002Fmo):** Increased cloud quota + unlimited local models\n- **AI Ultimate ($25\u002Fmo):** Maximum cloud quota + advanced features\n\n**[AI Pricing](https:\u002F\u002Fwww.jetbrains.com\u002Fai-ides\u002Fbuy\u002F)** | **[AI Features](https:\u002F\u002Fwww.jetbrains.com\u002Fai-assistant\u002F)**\n\n---\n\n#### [Tabnine](https:\u002F\u002Fwww.tabnine.com\u002F)\n\n**Models:** Claude 3.5 Sonnet, GPT-4o, Llama 3.3 70B, proprietary models\n- Free tier with limited features\n- Basic AI code completions and chat (limited)\n- Local processing available\n- Context heavily limited in free tier\n- 600+ programming languages supported\n- **Pro ($12\u002Fmo):** Enhanced AI completions and chat\n- **Enterprise ($39\u002Fuser\u002Fmo):** Multiple LLMs, private deployment, on-premises and air-gapped options\n\n**[Pricing](https:\u002F\u002Fwww.tabnine.com\u002Fpricing\u002F)**\n\n---\n\n#### [SuperMaven](https:\u002F\u002Fsupermaven.com\u002F) ⚠️ DISCONTINUED\n\n**Status:** Shut down November 21, 2025 after acquisition by Cursor (Nov 2024)\n\n~~**Models:** GPT-4o, Claude 3.5 Sonnet, GPT-4 (via chat interface)~~\n~~- Free tier with basic features~~\n~~- Basic code suggestions~~\n~~- 7-day data retention limit~~\n~~- Credit card required for registration~~\n~~- 1M token context window~~\n\n**Historical Note:** SuperMaven was acquired by Cursor in November 2024 and officially shut down in November 2025. Features were integrated into Cursor Tab. Users should migrate to Cursor or alternatives.\n\n**[Pricing](https:\u002F\u002Fsupermaven.com\u002Fpricing)**\n\n---\n\n#### [Bolt.new](https:\u002F\u002Fbolt.new\u002F)\n\n**Models:** Unspecified models\n- **$1 credit\u002Fmo = ~100K tokens** (reduced Mar 2026)\n- Specific model not publicly specified\n- Credit card required\n- **$20\u002Fmo:** 20M tokens\u002Fmonth\n- **$200\u002Fmo:** 200M tokens\u002Fmonth\n\n**[Token Documentation](https:\u002F\u002Fsupport.bolt.new\u002Faccount-and-subscription\u002Ftokens)**\n\n---\n\n#### [Lovable](https:\u002F\u002Flovable.dev\u002F)\n\n**Models:** Unspecified models\n- 5 daily credits, max 30 per month (free)\n- Models not publicly enumerated\n- Credit card required\n- **Pro ($25\u002Fmo):** 150 credits\u002Fmonth (5 daily credits)\n- **Teams ($30\u002Fmo):** Higher limits (undisclosed)\n\n**[Messaging Limits](https:\u002F\u002Fdocs.lovable.dev\u002Fuser-guides\u002Fmessaging-limits)**\n\n---\n\n#### [v0.dev](https:\u002F\u002Fv0.dev\u002F)\n\n**Models:** Proprietary models (not frontier)\n- $5 in credits\u002Fmonth limit\n- Uses proprietary models with varied routing\n- Credit card required\n- GPT-5 access requires v0 Premium subscription\n\n**[Updated Pricing Blog](https:\u002F\u002Fvercel.com\u002Fblog\u002Fimproved-v0-pricing-5luSrdRUJsRvf1kXWoYGxh)**\n\n---\n\n### Additional 2026 AI Chat Platforms\n\nGeneral-purpose chat interfaces with free tiers.\n\n| Platform | Free Model | Key Capabilities | Limitations |\n|----------|------------|------------------|-------------|\n| [ChatGPT](https:\u002F\u002Fchatgpt.com) | **GPT-4o \u002F GPT-5.4-limited** [verify] | Sora 3, DALL-E 4, GPT Store | ~20 msgs\u002F3hr |\n| [Gemini](https:\u002F\u002Fgemini.google.com) | **Gemini 3.1 Flash** | 2M Context, **20 Deep Research\u002Fmo** | Research quota |\n| [Claude](https:\u002F\u002Fclaude.ai) | **Claude Sonnet\u002FHaiku** [verify: Opus paid-only] | Technical reasoning | ~30 msgs\u002F5h |\n| [Grok](https:\u002F\u002Fgrok.com) | **Grok 4.2** | Aurora 2 images, voice | 15 msgs\u002F12hr |\n| [Mistral Le Chat](https:\u002F\u002Fchat.mistral.ai) | **Mistral Medium 3** | Structured output | Fewer integrations |\n\n**Notes:**\n- **Aurora** - xAI's image generation model (available in Grok)\n- **Sora 2** - OpenAI's video generation (integrated in ChatGPT)\n- **DALL-E 4** - OpenAI's latest image model (ChatGPT)\n- **Deep Research** - Gemini's agentic research feature\n\n---\n\n## CLI Coding Tools\n\nCommand-line tools for AI-assisted coding in your terminal.\n\n### CLI Tools with Pro-Grade Models\n\n#### [Gemini CLI](https:\u002F\u002Fgithub.com\u002Fgoogle-gemini\u002Fgemini-cli)\n\n**Models:** Gemini 3.1 Flash [verify: Pro now paid], Gemini 2.5 Pro\n- Gemini 3.1 Pro latest version (v0.37.1 April 2026)\n- 100 requests\u002Fday for Gemini 2.5 Pro (free tier fallback)\n- 250 requests\u002Fday for Gemini 2.5 Flash\n- No credit card required for free tier\n- MCP server support, Google Search grounding\n- Enable via `\u002Fsettings` → Preview features → true\n- **Install:** `npm install -g @google\u002Fgemini-cli`\n\n**[Rate Limits](https:\u002F\u002Fai.google.dev\u002Fgemini-api\u002Fdocs\u002Frate-limits)** | **[Pricing](https:\u002F\u002Fai.google.dev\u002Fgemini-api\u002Fdocs\u002Fpricing)** | **[Gemini 3 Pro Announcement](https:\u002F\u002Fdevelopers.googleblog.com\u002Fen\u002F5-things-to-try-with-gemini-3-pro-in-gemini-cli\u002F)**\n\n---\n\n#### [Rovo Dev CLI](https:\u002F\u002Fwww.atlassian.com\u002Fblog\u002Fannouncements\u002Frovo-dev-command-line-interface)\n\n> [!IMPORTANT]  \n> Rovo Dev CLI isn’t available during a Rovo Dev Standard trial. To use this feature, you need a paid Rovo Dev Standard subscription.\n\n**Models:** Claude Sonnet 4 [verify], GPT-5 preview [verify]\n- 5M tokens\u002Fday free tier\n- No credit card required during beta\n- Token limits reset at midnight UTC\n- Jira\u002FConfluence integration, MCP server support\n- Requires Atlassian account\n- **Pro ($19.99\u002Fmo via Google AI Pro):** 100 tasks\u002Fday, 5x higher limits, 5x concurrent tasks (15)\n- **Ultra (via Google AI Ultra):** 300 tasks\u002Fday, 20x higher limits, 60 concurrent tasks, priority access to latest models\n\n**[Documentation](https:\u002F\u002Fsupport.atlassian.com\u002Frovo\u002Fdocs\u002Fuse-rovo-dev-cli\u002F)** | **[Token Limits](https:\u002F\u002Fsupport.atlassian.com\u002Frovo\u002Fdocs\u002Frovo-dev-cli-limits\u002F)**\n\n---\n\n#### [Warp](https:\u002F\u002Fwarp.dev\u002F)\n\n**Models:** GPT-4.1, Claude Opus 4.1 [verify], Claude Sonnet 4 [verify], Gemini 2.5 Pro\n- 150 AI credits\u002Fmonth (first 2 months), then 75 AI credits\u002Fmonth\n- No credit card required for basic signup\n- AI-powered terminal with code generation\n- **Build ($20\u002Fmo):** 1,500 AI credits\u002Fmonth\n- Reload Credits available (up to 50% cheaper than old overage rates, roll over for 12 months)\n- Bring Your Own API Key (BYOK) option available\n- New pricing effective immediately for new customers (Oct 30, 2025)\n- Existing monthly subscribers transition on first renewal after Dec 1, 2025\n\n**[Pricing](https:\u002F\u002Fwww.warp.dev\u002Fpricing)**\n\n---\n\n#### [GitHub Copilot](https:\u002F\u002Fgithub.com\u002Ffeatures\u002Fcopilot\u002Fplans)\n\n**Models:** GPT-4.1, Claude Opus 3.5, Gemini 2.0 Flash, Grok Code Fast 1 (Free tier); GPT-5.1-Codex-Max available in Pro\u002FPro+\u002FBusiness\u002FEnterprise only\n- 50 agent mode or chat requests + 2,000 completions\u002Fmonth (Free tier)\n- Agent Mode with autonomous multi-step coding\n- No credit card required\n- Free Copilot Pro for students\u002Feducators (GitHub Student Pack, Copilot Pro for teachers\u002Fmaintainers)\n- Limited to basic features after quota\n- **Pro ($10\u002Fmo):** 300 premium requests + unlimited completions\u002Fmonth\n- **Pro+ ($39\u002Fmo):** 1,500 premium requests + unlimited completions\u002Fmonth\n- **Business ($19\u002Fuser\u002Fmo):** 300 premium requests\u002Fuser + unlimited completions\n- **Enterprise ($39\u002Fuser\u002Fmo):** 1,000 premium requests\u002Fuser + unlimited completions\n- **GPT-5.1-Codex-Max** available in public preview (Dec 4, 2025) for Pro, Pro+, Business, Enterprise - **NOT in free tier**\n- Overage billing available at $0.04\u002Frequest\n\n**[Plans Details](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fcopilot\u002Fget-started\u002Fplans-for-github-copilot)** | **[Agent Mode](https:\u002F\u002Fcode.visualstudio.com\u002Fblogs\u002F2025\u002F02\u002F24\u002Fintroducing-copilot-agent-mode)** | **[GPT-5.1-Codex-Max Preview](https:\u002F\u002Fgithub.blog\u002Fchangelog\u002F2025-12-04-openais-gpt-5-1-codex-max-is-now-in-public-preview-for-github-copilot\u002F)**\n\n---\n\n#### [Jules](https:\u002F\u002Fjules.google\u002F)\n\n**Model:** Gemini 2.5 Pro\n- 15 tasks\u002Fday free tier\n- 3 concurrent tasks\n- No credit card required\n- Gmail account required (18+ years)\n- Task limits reset on rolling 24-hour window\n- **Pro ($19.99\u002Fmo):** 100 tasks\u002Fday, 5x higher limits, 5x concurrent tasks (15)\n- **Ultra (via Google AI Ultra):** 300 tasks\u002Fday, 20x higher limits, 60 concurrent tasks, priority access to latest models\n\n**[Usage Limits](https:\u002F\u002Fjules.google\u002Fdocs\u002Fusage-limits\u002F)** | **[Documentation](https:\u002F\u002Fjules.google\u002Fdocs\u002F)** | **[Google AI Plans](https:\u002F\u002Fone.google.com\u002Fabout\u002Fgoogle-ai-plans\u002F)**\n\n---\n\n#### [AWS Kiro](https:\u002F\u002Fkiro.dev\u002F)\n\n**Models:** Claude 4 Sonnet, Claude 3.7 Sonnet (AWS-hosted)\n- 50 credits\u002Fmonth (Free tier)\n- 14-day welcome bonus: 500 credits\n- No credit card required\n- **Pro ($20\u002Fmo):** 1,000 credits\n- **Pro+ ($40\u002Fmo):** 2,000 credits\n- **Power ($200\u002Fmo):** 10,000 credits\n\n**[Pricing](https:\u002F\u002Fkiro.dev\u002Fpricing\u002F)** | **[Introduction Blog](https:\u002F\u002Fkiro.dev\u002Fblog\u002Fintroducing-kiro\u002F)**\n\n---\n\n#### [Amazon Q Developer](https:\u002F\u002Faws.amazon.com\u002Fq\u002Fdeveloper\u002F)\n\n**Model:** Claude Sonnet 4 [verify] (AWS-hosted)\n- 50 agentic requests\u002Fmonth limit (multi-turn conversations)\n- Latest Claude models\n- Credit card required\n- Must upgrade to Pro for continued access\n- Perpetual free tier\n- **Pro ($19\u002Fmo):** Increased limits for agentic requests\n- Usage may be adjusted based on regional factors and usage patterns\n\n**[Pricing](https:\u002F\u002Faws.amazon.com\u002Fq\u002Fdeveloper\u002Fpricing\u002F)**\n\n---\n\n#### [OpenCode](https:\u002F\u002Fgithub.com\u002Fanomalyco\u002Fopencode)\n\n**Models:** 300+ via OpenRouter (Claude, GPT, DeepSeek, Gemini, Grok, etc.)\n- Open-source AI coding agent (Go-based CLI)\n- Zen Free tier with 8 exclusive models (Big Pickle, MiniMax M2.5 Free, MiMo V2)\n- Privacy-sensitive: no code\u002Fcontext stored\n- `opencode run --dangerously-skip-perm` for quick execution\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002Fanomalyco\u002Fopencode)** | **[Website](https:\u002F\u002Fopencode.ai)**\n\n---\n\n#### [ForgeCode](https:\u002F\u002Fgithub.com\u002Fantinomyhq\u002Fforgecode)\n\n**Models:** 300+ models via OpenRouter (Claude, GPT, O Series, Grok, DeepSeek, Gemini)\n- AI-enabled pair programmer (Rust-based, Apache 2.0)\n- Model-agnostic agent harness\n- Semantic codebase search via `:sync`\n- 10K tokens\u002Fday free tier\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002Fantinomyhq\u002Fforgecode)** | **[Website](https:\u002F\u002Fforgecode.dev)**\n\n---\n\n#### [OhMyPi](https:\u002F\u002Fgithub.com\u002Fcan1357\u002Foh-my-pi)\n\n**Models:** Bring your own keys (any provider)\n- AI coding agent for the terminal (Zig-powered)\n- Hash-anchored edits, optimized tool harness\n- LSP integration, Python support, browser automation\n- Subagents with coordinated API rate limiting\n- Multiplexer integration (tmux, GNU Screen, Zellij)\n- Interrupt anytime workflow\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002Fcan1357\u002Foh-my-pi)**\n\n---\n\n#### [Goose](https:\u002F\u002Fgithub.com\u002Fblock\u002Fgoose)\n\n**Models:** Any LLM (Claude, GPT, DeepSeek, etc.)\n- Open-source extensible AI agent from Block (now AAIF\u002FLinux Foundation)\n- Desktop app, CLI, and API\n- Active engineering tasks (not just code suggestions)\n- Built for code, workflows, and automation\n- Model-agnostic architecture\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002Fblock\u002Fgoose)** | **[Website](https:\u002F\u002Fblock.github.io\u002Fgoose\u002F)**\n\n---\n\n#### [Kilo Code](https:\u002F\u002Fkilocode.ai\u002F)\n\n**Models:** Bring your own API keys (Claude, Gemini, GPT, etc.)\n- Up to $25 signup credits (one-time bonus)\n- Open source VS Code extension\n- Pay-as-you-go with no markup on model pricing\n- Credit card required to claim full bonus credits\n- Full BYOK support\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002FKilo-Org\u002Fkilocode)** | **[Documentation](https:\u002F\u002Fkilocode.ai\u002Fdocs\u002F)** | **[Pricing](https:\u002F\u002Fkilocode.ai\u002Fpricing)**\n\n---\n\n#### [RooCode](https:\u002F\u002Fgithub.com\u002FRooCodeInc\u002FRoo-Code)\n\n**Models:** Bring your own API keys (any provider)\n- Open-source AI-powered coding assistant for VS Code\n- Whole dev team of AI agents in your editor\n- No subscription required - pay-as-you-go with your own keys\n- Custom modes for different coding tasks\n- Previously known as Roo Cline\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002FRooCodeInc\u002FRoo-Code)** | **[Website](https:\u002F\u002Froocode.com)**\n\n---\n\n### CLI Tools with Basic Models\n\n#### [Claude Code](https:\u002F\u002Fwww.anthropic.com\u002Fclaude-code)\n\n**Models:** Claude Sonnet 4 [verify], Opus 4.5 [verify: paid-only], Haiku 4.5\n- Free tier available with limited usage\n- Pro ($20\u002Fmo or $17\u002Fmo annually): Sonnet 4 access with more usage\n- Max 5x ($100\u002Fmo): ~225 messages\u002F5 hours\n- Max 20x ($200\u002Fmo): ~900 messages\u002F5 hours\n- Extended thinking modes: \"think\" (~4K tokens), \"megathink\" (~10K), \"ultrathink\" (~32K)\n- Usage limits reset weekly with 5-hour rolling windows\n\n**[Pricing](https:\u002F\u002Fwww.anthropic.com\u002Fpricing)** | **[Claude Code Guide](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code)**\n\n---\n\n#### [OpenAI Codex CLI](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fcodex)\n\n**Model:** GPT-5.1-Codex-Max (77.9% SWE-bench Verified)\n- Free with ChatGPT Plus ($20\u002Fmo): 30–150 messages\u002F5 hours\n- ChatGPT Pro ($200\u002Fmo): 300–1,500 messages\u002F5 hours\n- Pay-as-you-go API: $1.25\u002F$10 per million tokens (input\u002Foutput)\n- Free OSS mode: Access to open-source models only (via `--oss` flag)\n- First model with \"compaction\" for multi-million token sessions (24+ hour tasks)\n- 30% fewer thinking tokens than previous GPT-5.1-Codex\n- Cross-platform: macOS 12+, Ubuntu 20.04+, Windows 11 via WSL2\n\n**[GitHub Repo](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fcodex)** | **[GPT-5.1-Codex-Max Announcement](https:\u002F\u002Fopenai.com\u002Findex\u002Fgpt-5-1-codex-max\u002F)**\n\n---\n\n#### [QuantFlow Pilot](https:\u002F\u002Fgithub.com\u002Fqf-studio\u002Fpilot)\n\n**Models:** Uses Claude Code for implementation\n- Autonomous AI development pipeline — #1 Terminal Benchmark 2.0\n- Turns GitHub issues into pull requests automatically\n- Label an issue \"pilot\" → Pilot claims it → Creates branch → Plans → Implements → Quality gates → Opens PR\n- Telegram bot integration available\n- Desktop app available\n- **Install:** `brew install qf-studio\u002Ftap\u002Fpilot` or `go install github.com\u002Fqf-studio\u002Fpilot@latest`\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002Fqf-studio\u002Fpilot)** | **[Website](https:\u002F\u002Fpilot.quantflow.studio\u002F)**\n\n---\n\n#### [MemoryPalace](https:\u002F\u002Fgithub.com\u002Fmilla-jovovich\u002Fmempalace)\n\n**Models:** Works with any LLM (Claude, ChatGPT, Cursor, Gemini, local models)\n- AI memory system with highest LongMemEval score ever (96.6%)\n- Uses ancient \"memory palace\" technique for AI conversations\n- Stores conversations in structured format: wings (people\u002Fprojects), halls (memory types), rooms (specific ideas)\n- Raw verbatim storage without AI summarization\n- Three mining modes: projects (code\u002Fdocs), convos (conversation exports), general (auto-classified)\n- MCP server with 19 tools for AI integration\n- Local, open, adaptable — runs entirely on your machine\n- **Install:** `pip install mempalace`\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002Fmilla-jovovich\u002Fmempalace)** | **[Note from Milla](https:\u002F\u002Fgithub.com\u002Fmilla-jovovich\u002Fmempalace#readme)**\n\n---\n\n#### [Continue.dev](https:\u002F\u002Fwww.continue.dev\u002F)\n\n**Models:** Bring your own API keys (200+ models supported)\n- Free VS Code and JetBrains extension\n- Full support for local models via Ollama\u002FLM Studio\n- Solo tier: Private\u002Fteam\u002Fpublic visibility options\n- Community hub for custom AI assistants\n- No vendor lock-in or usage limits for local models\n\n**[GitHub](https:\u002F\u002Fgithub.com\u002Fcontinuedev\u002Fcontinue)** | **[Model Hub](https:\u002F\u002Fhub.continue.dev\u002Fexplore\u002Fmodels)**\n\n---\n\n#### [Aider](https:\u002F\u002Faider.chat\u002F)\n\n**Models:** Bring your own API keys (supports many providers)\n- Free command-line assistant with built-in Git integration\n- Works with GPT-4o, Claude Sonnet, DeepSeek, and local models\n- Multi-file editing with repository context\n- Voice-to-code support\n- Use `\u002Fhelp` to see all commands\n\n**[Documentation](https:\u002F\u002Faider.chat\u002F)** | **[GitHub](https:\u002F\u002Fgithub.com\u002Fpaul-gauthier\u002Faider)**\n\n---\n\n## API Providers for AI Coding Tools\n\nThese services provide API access to coding-optimized models for tools like Cursor, Continue.dev, Cline, etc.\n\n### [OpenRouter](https:\u002F\u002Fopenrouter.ai\u002F)\n\n- 50 requests\u002Fday free tier (1,000\u002Fday with $10+ credits)\n- Qwen3-Coder-480B, Qwen3-30B-A3B, Qwen3-235B-A22B, Gemini Flash\n- 20 req\u002Fmin rate limit for free tier\n- OpenAI-compatible API\n\n### [Cerebras](https:\u002F\u002Fcloud.cerebras.ai\u002F)\n\n- **1.5M tokens\u002Fday** free tier (expanded Feb 2026)\n- 30 req\u002Fmin, 8,192 token context\n- Models: **Qwen3.6-Plus-480B**, Llama 3.1 70B\n- Ultra-fast: **2,400 t\u002Fs** (Qwen3.6)\n- OpenAI-compatible API (works with Cursor, Continue.dev, Cline, RooCode, etc.)\n- **Paid tiers:** Developer ($10+ self-serve), Enterprise (custom pricing)\n\n**[Pricing](https:\u002F\u002Fwww.cerebras.ai\u002Fpricing)** | **[API Docs](https:\u002F\u002Finference-docs.cerebras.ai\u002F)** | **[Integrations](https:\u002F\u002Finference-docs.cerebras.ai\u002Fintegrations\u002F)**\n\n---\n\n## Paid Tiers Comparison\n\n### AI-Powered IDEs - Paid Plans\n\n| IDE | Entry Tier | Credits\u002FRequests | Key Features |\n|-----|------------|------------------|--------------|\n| [Cursor](https:\u002F\u002Fcursor.com\u002F) | Pro ($20\u002Fmo) | Extended Agent limits | Unlimited completions |\n| [Trae](https:\u002F\u002Ftrae.ai\u002F) | Pro ($10\u002Fmo) | 600 fast + unlimited slow | Zero rate limits |\n| [Windsurf](https:\u002F\u002Fwindsurf.com\u002F) | Pro ($20\u002Fmo) | 500 prompt credits | Multi-provider |\n| [Qoder](https:\u002F\u002Fqoder.com\u002F) | Pro ($10\u002Fmo - 50% off) | 2,000 credits | Quest Mode, Experts Mode |\n| [Codeium](https:\u002F\u002Fcodeium.com\u002F) | Pro ($10\u002Fmo) | Unlimited | Claude 4.6 [verify], GPT-5.4 [verify] |\n| [Tabnine](https:\u002F\u002Fwww.tabnine.com\u002F) | Pro ($12\u002Fmo) | Enhanced completions | 600+ languages |\n| [JetBrains AI](https:\u002F\u002Fwww.jetbrains.com\u002Fai\u002F) | AI Pro ($15\u002Fmo) | Increased cloud quota | Unlimited local models |\n| ~~[SuperMaven](https:\u002F\u002Fsupermaven.com\u002F)~~ | ~~DISCONTINUED Nov 2025~~ | ~~1M token context~~ | ~~Acquired by Cursor~~ |\n\n### CLI Tools - Paid Plans\n\n| Tool | Entry Tier | Credits\u002FRequests | Key Features |\n|------|------------|------------------|--------------|\n| [Claude Code](https:\u002F\u002Fwww.anthropic.com\u002Fclaude-code) | Pro ($20\u002Fmo) | ~225 messages\u002F5h | Sonnet access [verify] |\n| [Warp](https:\u002F\u002Fwarp.dev\u002F) | Build ($20\u002Fmo) | 1,500 credits\u002Fmonth | BYOK available |\n| [GitHub Copilot](https:\u002F\u002Fgithub.com\u002Ffeatures\u002Fcopilot) | Pro ($10\u002Fmo) | 300 premium req\u002Fmonth | Unlimited completions |\n| [Rovo Dev CLI](https:\u002F\u002Fwww.atlassian.com\u002Frovo) | Jira Standard ($7.53\u002Fmo) | 20M tokens\u002Fday | 4x free tier |\n| [Jules](https:\u002F\u002Fjules.google\u002F) | Pro ($19.99\u002Fmo) | 100 tasks\u002Fday | 5x free limits |\n| [OpenAI Codex CLI](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fcodex) | ChatGPT Plus ($20\u002Fmo) | 30-150 msg\u002F5h | GPT-5.1-Codex-Max |\n| [Amazon Q Developer](https:\u002F\u002Faws.amazon.com\u002Fq\u002Fdeveloper\u002F) | Pro ($19\u002Fmo) | Increased agentic limits | AWS-hosted Claude |\n| [Kilo Code](https:\u002F\u002Fkilocode.ai\u002F) | Pay-as-you-go | Up to $25 signup credits | No markup on models |\n\n---\n\n## Local Models\n\nRunning open-weight frontier models locally provides unlimited coding assistance without API costs.\n\n**Popular Tools:**\n- **[Cline](https:\u002F\u002Fcline.bot\u002F)** - VS Code extension with Plan\u002FAct modes and MCP support\n- **[Aider](https:\u002F\u002Faider.chat\u002F)** - Command-line assistant with Git integration\n- **[Continue.dev](https:\u002F\u002Fwww.continue.dev\u002F)** - Open-source VS Code extension (200+ models)\n\n**Local Model Tools:**\n- **[Ollama](https:\u002F\u002Follama.com\u002F)** - Run frontier models locally\n- **[LM Studio](https:\u002F\u002Flmstudio.ai\u002F)** - Easy desktop app for local LLMs (no terminal required)\n\n**Notable Local Models (2026):**\n\n- Qwen3.6-Plus-480B (71.2% SWE, ~150GB VRAM)\n- **Gemma 4** [verify] (Google, Apache 2.0, fully open-source)\n- **GLM-5.1 \u002F GLM-5V-Turbo** [verify] (Zhipu MoE-based SOTA coders)\n- Devstral 2 (24B, Apache 2.0, agent-optimized)\n- DeepSeek Coder V4 (lite version ~18GB)\n- Codestral 2 (Mistral, 22B)\n- GLM-4.9-Air (Chinese\u002FEnglish coding)\n\n> **Note:** Frontier models require substantial RAM\u002FVRAM. See [Unsloth Qwen3-Coder guide](https:\u002F\u002Fdocs.unsloth.ai\u002Fbasics\u002Fqwen3-coder-how-to-run-locally) for details.\n\n> **Update April 2026:** Gemma 4 and GLM-5.1 families are new flagship open-source releases. Verify availability in Ollama\u002FLM Studio before downloading.\n\n---\n\n## free-coding-models CLI\n\nFind the fastest free coding model in seconds. Ping 238 models across 25 providers in real-time.\n\n```bash\nnpm install -g free-coding-models\nfree-coding-models\n```\n\n### Features\n\n- **Parallel pings** — all 238 models tested simultaneously\n- **Stability Score (0-100)** — composite score from p95 latency, jitter, spike rate, uptime\n- **Smart ranking** — top 3 highlighted 🥇🥈🥉\n- **Favorites** — star models with `F`, persisted across sessions\n- **Tool Integration** — auto-configure OpenCode, Goose, Aider, Continue, Cline, etc.\n- **OpenCode Zen Models** — 8 exclusive free models (Big Pickle, MiniMax M2.5 Free, MiMo V2, etc.)\n\n### Quick Usage\n\n```bash\n# Most reliable model right now\nfree-coding-models --fiable\n\n# Configure Goose with S-tier model\nfree-coding-models --goose --tier S\n\n# NVIDIA top models only\nfree-coding-models --origin nvidia --tier S\n\n# JSON output for scripting\nfree-coding-models --tier S --json | jq -r '.[0].modelId'\n```\n\n### Tool Launcher Flags\n\n| Flag | Launches |\n|------|----------|\n| `--opencode` | 📦 OpenCode CLI |\n| `--openclaw` | 🦞 OpenClaw |\n| `--goose` | 🪿 Goose |\n| `--aider` | 🛠 Aider |\n| `--qwen` | 🐉 Qwen Code |\n| `--continue` | ▶️ Continue CLI |\n| `--cline` | 🧠 Cline |\n| `--gemini` | ♊ Gemini CLI |\n| `--rovo` | 🦘 Rovo Dev CLI |\n| And 8 more... |\n\n### Tier Scale\n\n| Tier | SWE-bench | Best For |\n|------|-----------|----------|\n| **S+** | ≥75% | **Claude Opus 4.6 [verify], GPT-5.4 [verify]** |\n| **S** | 65-75% | **Qwen3.6-Plus (71.2%), Claude Sonnet 4.6 [verify]** |\n| **A+\u002FA** | 40–60% | Solid alternatives |\n| **A-\u002FB+** | 30–40% | Smaller tasks |\n| **B\u002FC** | \u003C 30% | Code completion |\n\n### License Summary\n\nAll 238 models allow **commercial use of generated output**. You own what the models generate.\n\n| License | Models | Commercial |\n|---------|--------|:----------:|\n| Apache 2.0 | Qwen3\u002FQwen2.5 Coder, GPT-OSS 120B\u002F20B, Devstral Small 2, Gemma 4, MiMo V2 Flash | ✅ Unrestricted |\n| MIT | GLM 4.5\u002F4.6\u002F4.7\u002F5, MiniMax M2.1, Devstral 2 | ✅ Unrestricted |\n| Llama Community License | Llama 3.3 70B, Llama 4 Scout\u002FMaverick | ✅ Attribution required. >700M MAU → separate Meta license |\n| DeepSeek License | DeepSeek V3\u002FV3.1\u002FV3.2, R1 | ✅ Use restrictions on model (no military, no harm) — output is yours |\n| NVIDIA Nemotron License | Nemotron Super\u002FUltra\u002FNano | ✅ Updated Mar 2026, now near-Apache 2.0 permissive |\n| MiniMax Model License | MiniMax M2, M2.5 | ✅ Royalty-free, non-exclusive. Prohibited uses policy applies to model |\n| Proprietary (API) | Claude (Rovo), Gemini (CLI), Perplexity Sonar, Mistral Large, Codestral | ✅ You own outputs per provider ToS |\n| OpenCode Zen | Big Pickle, MiMo V2 Pro\u002FFlash\u002FOmni Free, GPT 5 Nano, MiniMax M2.5 Free, Nemotron 3 Super Free | ✅ Per OpenCode Zen ToS |\n\n**Key Points:**\n1. **Generated code is yours** — no model claims ownership of your output\n2. **Apache 2.0 \u002F MIT models** (Qwen, GLM, GPT-OSS, MiMo, Devstral Small) are the most permissive — no strings attached\n3. **Llama** requires \"Built with Llama\" attribution; >700M MAU needs a Meta license\n4. **DeepSeek \u002F MiniMax** have use-restriction policies (no military use) that govern the model, not your generated code\n5. **API-served models** (Claude, Gemini, Perplexity) grant full output ownership under their terms of service\n\n> ⚠️ **Disclaimer:** This is a summary, not legal advice. License terms can change. Always verify the current license on the model's official page before making legal decisions.\n\n---\n\n## Comparison Notes\n\n- **Goal**: Compare AI coding tools by their access to pro-grade models and free tier limits.\n- **What qualifies a model as \"pro-grade\"?** Models must achieve ≥60% on SWE-bench Verified, demonstrating real-world software engineering capability. Current qualifying models: Claude Opus 4.5 (80.9% [verify]), GPT-5.1-Codex-Max (77.9% [verify]), Claude Sonnet 4.5 (77.2% [verify]), Gemini 3 Pro (76.2% [verify]), GPT-5 (74.9% [verify]), Claude Opus 4.1 (74.5% [verify]), Claude Sonnet 4 (72.7% [verify]), GPT-5 mini (71.0% [verify]), Qwen3-Coder-480B (69.6% [verify]), and Gemini 2.5 Pro (63.2% [verify]).\n- **`[verify]` tag**: Indicates information needs verification from official sources. Pricing, limits, and model availability change frequently.\n- **Different limit types**: Tools use various quota systems - requests, tokens, credits, chats - making direct comparison challenging. Check documentation for specifics.\n- **Real-world usage**: Actual consumption varies dramatically based on coding style, task complexity, and tool implementation.\n\n---\n\n## Education & Student Programs\n\n| Program | What You Get | Requirements |\n|---------|--------------|--------------|\n| [GitHub Student Pack](https:\u002F\u002Feducation.github.com\u002Fpack) | Free Copilot Pro for students | Verify with .edu email |\n| [GitHub Copilot Free](https:\u002F\u002Fcode.visualstudio.com\u002Fblogs\u002F2024\u002F12\u002F18\u002Ffree-github-copilot) | 50 chat + 2,000 completions\u002Fmonth | VS Code users |\n| [Copilot Pro for Teachers\u002FMaintainers](https:\u002F\u002Fdocs.github.com\u002Fen\u002Fcopilot\u002Fhow-tos\u002Fmanage-your-account\u002Fget-free-access-to-copilot-pro) | Free Copilot Pro | Open source maintainers & educators |\n\n---\n\n## Additional 2026 AI Tools\n\n### Agentic Workflow Platforms\n\nVisual orchestration tools for building autonomous AI agents without coding.\n\n| Platform | Free Tier | Best For | Key Features |\n|----------|-----------|----------|--------------|\n| [Make](https:\u002F\u002Fmake.com) (Integromat) | 1,000 ops\u002Fmonth | Visual builders | Drag-and-drop AI Agents, 3,000+ app integrations |\n| [n8n](https:\u002F\u002Fn8n.io) | Unlimited (self-hosted) | Technical teams | Self-hosted RAG systems, private data automation |\n| [Gumloop](https:\u002F\u002Fgumloop.com) | 2,000 credits\u002Fmonth | No-code agents | Natural-language builder, \"Gummie\" troubleshooting agent |\n| [Relay.app](https:\u002F\u002Frelay.app) | Generous free plan | Beginners | Simple agentic workflows |\n| [Activepieces](https:\u002F\u002Factivepieces.com) | 1,000 tasks\u002Fmonth | Open-source | Flat pricing, self-hostable |\n| [Podium](https:\u002F\u002Fpodium.com) | Entry-level tiers | Sales\u002Fcommunication | 24\u002F7 lead response AI agents |\n| [QuantFlow Pilot](https:\u002F\u002Fgithub.com\u002Fqf-studio\u002Fpilot) | Free | Autonomous development | #1 Terminal Benchmark 2.0 — AI that ships your tickets |\n\n---\n\n### Data Visualization & Analysis\n\nAI-powered tools for conversational data analysis and narrative visualization.\n\n| Tool | Function | Free Tier Detail | Key Feature |\n|------|----------|------------------|-------------|\n| [Julius](https:\u002F\u002Fjulius.ai) | Chat-with-data | Upload spreadsheets, generate instant visualizations |\n| [Anomaly AI](https:\u002F\u002Ffindanomaly.ai) | AI Dashboards | Generate interactive dashboards from natural language |\n| [Flourish](https:\u002F\u002Fflourish.studio) | Data Storytelling | No-code interactive maps, \"scrollytelling\" features |\n| [Datawrapper](https:\u002F\u002Fdatawrapper.de) | Publishing | Publish-ready charts in seconds, journalism-focused |\n| [Looker Studio](https:\u002F\u002Flookerstudio.google.com) | Marketing Data | Seamless Google Analytics\u002FAds integration |\n| [Power BI Desktop](https:\u002F\u002Fpowerbi.microsoft.com) | Microsoft reports | Copilot recommendations, local report building |\n\n---\n\n### Creative & Multimedia Tools\n\nProfessional-grade content creation with generous free tiers.\n\n| Tool | Output | Free Tier | Key Capability |\n|------|--------|-----------|----------------|\n| [Veo](https:\u002F\u002Fdeepmind.google\u002Ftechnologies\u002Fveo\u002F) | Video | Basic Free | Cinematic clips with realistic motion and sound |\n| [Sora 2](https:\u002F\u002Fopenai.com\u002Fsora) (via ChatGPT) | Video | Limited free tier | Deep ChatGPT integration, high-quality video |\n| [DALL-E 4](https:\u002F\u002Fopenai.com\u002Fdall-e-4) (via ChatGPT) | Image | Limited free tier | Latest OpenAI image model |\n| [Synthesia](https:\u002F\u002Fsynthesia.io) | Video Avatars | Free individual plan | \"Video Agents\" in 120+ languages |\n| [1 More Shot](https:\u002F\u002Fonemoreshot.ai) | Music Videos | Free plan | Advanced lip-sync, frame-by-frame control |\n| [Leonardo.Ai](https:\u002F\u002Fleonardo.ai) | Images | 150 tokens\u002Fday (~70 images) | Commercial use allowed |\n| [Recraft AI](https:\u002F\u002Frecraft.ai) | Vector\u002FSVG | 30 credits\u002Fday | Infinitely scalable icons and logos |\n| [Ideogram](https:\u002F\u002Fideogram.ai) | Images | 10-20 prompts\u002Fday | Perfect text rendering, \"Magic Prompt\" |\n| [Suno AI](https:\u002F\u002Fsuno.ai) | Music | 50 credits\u002Fday (~10 tracks) | Complete songs with vocals and instruments |\n| [ElevenLabs](https:\u002F\u002Felevenlabs.io) | Voice | Basic Free | Realistic voice cloning |\n| [Canva AI](https:\u002F\u002Fcanva.com) | Design | Robust free tier | AI design assets, brochures, short videos |\n\n---\n\n### Productivity & Research Tools\n\n| Tool | Function | Free Tier Detail | Key Feature |\n|------|----------|------------------|-------------|\n| [Grammarly](https:\u002F\u002Fgrammarly.com) | Writing | 100 AI prompts\u002Fmonth | Rewrites and tone detection |\n| [LanguageTool](https:\u002F\u002Flanguagetool.org) | Grammar | 10,000 characters\u002Ftext | 25+ languages, open-source |\n| [Fathom](https:\u002F\u002Ffathom.video) | Meetings | Forever Free | Records\u002Ftranscribes Zoom\u002FTeams, auto-sync to CRM |\n| [NotebookLM](https:\u002F\u002Fnotebooklm.google.com) | Research | Free | Audio Overview podcasts, grounded in your documents |\n| [Humata](https:\u002F\u002Fhumata.ai) | PDF Analysis | 60 pages\u002Fmonth | Clickable source citations |\n| [QuillBot](https:\u002F\u002Fquillbot.com) | Rewriting | 125 words\u002Ftime | Fluency & Standard modes |\n| [DeepL](https:\u002F\u002Fdeepl.com) | Translation | Basic Free | Incognito sensitive mode |\n| [MemoryPalace](https:\u002F\u002Fgithub.com\u002Fmilla-jovovich\u002Fmempalace) | AI Memory | Free, open source | 96.6% LongMemEval — memory palace technique for AI |\n\n---\n\n### Vertical AI (Specialized Domains)\n\n**Medical AI:**\n| Tool | Pricing | Key Value |\n|------|---------|-----------|\n| [iatroX](https:\u002F\u002Fiatrox.com) | Free | Adaptive Q-Bank, NICE\u002FBNF clinical reference |\n| [DxGPT](https:\u002F\u002Fdxgpt.com) | Free | Diagnostic assistant (500K+ users, 6K doctors) |\n| [OpenEvidence](https:\u002F\u002Fopenevidence.com) | Free (US verified) | Evidence-grounded search, ambient note generation |\n\n**Legal AI:**\n| Tool | Pricing | Key Value |\n|------|---------|-----------|\n| [DocLegal.Ai](https:\u002F\u002Fdoclegal.ai) | $10\u002Fmonth | Clause suggestion, risk detection |\n| [Doculex.ai](https:\u002F\u002Fdoculex.ai) | Varies | Case-data-driven drafting from medical records |\n| [Spellbook](https:\u002F\u002Fspellbook.legal) | 7-day trial | In-editor contract analysis |\n| [Harvey AI](https:\u002F\u002Fharvey.ai) | Enterprise | Regulatory matters, high security |\n\n---\n\n### Marketing & SEO Tools\n\n| Tool | Function |\n|------|----------|\n| [Wellows](https:\u002F\u002Fwellows.com) | AI Visibility Score tracking across ChatGPT, Gemini, Perplexity |\n| [Google SGE Labs](https:\u002F\u002Flabs.google.com) | See how AI Overviews interpret target keywords |\n| [NeuronWriter](https:\u002F\u002Fneuronwriter.com) | AI content scoring |\n| [Surfer SEO](https:\u002F\u002Fsurferseo.com) | Content optimization |\n| [Jasper](https:\u002F\u002Fjasper.ai) | AI copywriting with brand voice |\n| [Writesonic](https:\u002F\u002Fwritesonic.com) | Scalable copywriting |\n\n---\n\n### Open Source & Local Tools\n\n| Tool | Function | Description |\n|------|----------|-------------|\n| [Open WebUI](https:\u002F\u002Fopenwebui.com) | Local Chat Interface | ChatGPT-like experience running entirely offline with Ollama |\n| [Whisper](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fwhisper) (OpenAI) | Speech-to-Text | Most accurate open-source transcription |\n| [Piper](https:\u002F\u002Fgithub.com\u002Frhasspy\u002Fpiper) | Text-to-Speech | High-quality offline audio generation |\n| [ComfyUI](https:\u002F\u002Fcomfyui.org) | Image Generation | Node-based interface for Stable Diffusion |\n| [Zed](https:\u002F\u002Fzed.dev) | AI IDE | 50 AI prompts\u002Fmonth, native performance, high speed |\n| [Void IDE](https:\u002F\u002Fvoideditor.com\u002F) | Agent-first IDE | Multi-agent frontend\u002Fbackend\u002Ftesting | Preview, free tier |\n| [MemoryPalace](https:\u002F\u002Fgithub.com\u002Fmilla-jovovich\u002Fmempalace) | AI Memory System | 96.6% LongMemEval — memory palace technique for AI conversations | Free, open source |\n\n---\n\n## ⚡ Realtime & Streaming APIs\n\nLow-latency APIs for voice assistants, live coding copilots, trading tools, and realtime chat.\n\n### Streaming LLM APIs\n\n| Provider | Latency | Best For | Free Tier |\n|----------|---------|----------|-----------|\n| **Groq Streaming** | ~50-150ms (0.4ms\u002Ftoken) | Live coding, chat | 14.4K req\u002Fday |\n| **OpenAI Realtime API** | Low | Voice assistants, agents | **No free tier** (pay-per-use only, trial credits new accounts) |\n| **Gemini Live API** | Low | Multimodal streaming | **Dynamic caps** (varies by prompt complexity) |\n| **Cerebras** | **2,400 tok\u002Fsec** (Qwen3.6) | Batch + streaming | 1.5M tokens\u002Fday |\n| **Cloudflare Workers AI** | Edge | Global low-latency | 10K neurons\u002Fday |\n\n### Speech Streaming APIs\n\n| Provider | Type | Latency | Free Tier |\n|----------|------|---------|-----------|\n| **Deepgram** | STT streaming | ~300ms | $200 credits |\n| **AssemblyAI Streaming** | Realtime STT | ~400ms | 50 hours\u002Fmonth |\n| **Groq Whisper** | STT fast | ~200ms | 2,000 req\u002Fday |\n| **ElevenLabs Streaming** | TTS streaming | ~100ms | 10K chars\u002Fmonth |\n| **OpenAI Realtime** | STT + LLM + TTS | ~200ms | Limited |\n\n**Best for:**\n- **Trading bots:** Groq streaming (fastest)\n- **Voice assistants:** OpenAI Realtime API (end-to-end)\n- **Live captions:** AssemblyAI or Deepgram\n- **Realtime chat:** Gemini Live API\n\n---\n\n## 🎙️ Speech Models\n\nSpeech-to-text and text-to-speech models comparison.\n\n### Speech-to-Text (STT)\n\n| Model | Provider | Accuracy | Speed | Free Tier | Best For |\n|-------|----------|----------|-------|-----------|----------|\n| **Whisper Large v3** | OpenAI\u002FGroq\u002FLocal | Excellent | Fast | 2,000 req\u002Fday (Groq) | General purpose, local |\n| **Deepgram Nova** | Deepgram | Superior | Very Fast | $200 credits | Production, enterprise |\n| **AssemblyAI** | AssemblyAI | Excellent | Fast | 50 hours\u002Fmonth | Streaming, diarization |\n| **Whisper API** | OpenAI | Excellent | Medium | Pay-per-use | Reliable, consistent |\n| **Google Speech** | Google Cloud | Good | Fast | 60 min\u002Fmonth | Google ecosystem |\n| **Whisper (local)** | OpenAI\u002FOllama | Excellent | GPU-dependent | Unlimited offline | Privacy, cost control |\n\n### Text-to-Speech (TTS)\n\n| Model | Provider | Quality | Speed | Free Tier | Best For |\n|-------|----------|---------|-------|-----------|----------|\n| **ElevenLabs** | ElevenLabs | 🏆 Best | Fast | 10K chars\u002Fmonth | Voice cloning, pro voice |\n| **OpenAI TTS** | OpenAI | Excellent | Fast | Pay-per-use | Reliable, cheap |\n| **Piper** | Local | Good | Very Fast | Unlimited offline | Privacy, self-hosted |\n| **Bark** | Suno\u002FLocal | Good | Medium | Free (local) | Expressive, local |\n| **Google TTS** | Google Cloud | Good | Fast | 1M chars\u002Fmonth | Google ecosystem |\n| **WhisperSpeech** | Local | Good | Fast | Unlimited | Whisper-based TTS |\n\n### All-in-One Voice APIs\n\n| API | Input | Output | Latency | Use Case |\n|-----|-------|--------|---------|----------|\n| **OpenAI Realtime** | Audio | Audio | ~200ms | Voice agents |\n| **Deepgram Voice** | Audio | Text\u002FAudio | ~300ms | Voice bots |\n| **AssemblyAI LeMUR** | Audio | LLM response | ~1s | Voice RAG |\n\n---\n\n## 🎨 Image Generation Models\n\nComparison of image generation models and APIs.\n\n| Model | Provider | Quality | Speed | Free Tier | Best For |\n|-------|----------|---------|-------|-----------|----------|\n| **FLUX.2** | Black Forest Labs | 🏆 Excellent | Fast | Local\u002FReplicate | High quality, open |\n| **DALL-E 4** | OpenAI | 🏆 Best | Medium | ChatGPT Plus | Latest OpenAI |\n| **Ideogram 2.0** | Ideogram | Excellent | Fast | **20 prompts\u002Fday** | Text in images |\n| **Recraft V4** | Recraft | Excellent | Fast | **50 credits\u002Fday** | Vector\u002FSVG output |\n| **Stable Diffusion XL** | Stability AI | Good | Fast | Local\u002FDreamStudio | Flexibility, local |\n| **Midjourney v6** | Midjourney | 🏆 Excellent | Slow | None (paid only) | Artistic, Discord |\n| **Leonardo.ai** | Leonardo | Very Good | Fast | 150 tokens\u002Fday | Commercial use, gaming |\n| **Adobe Firefly** | Adobe | Good | Fast | 25 credits\u002Fmonth | Safe, commercial |\n| **Imagen 3** | Google | Excellent | Medium | Vertex AI trial | Photorealistic |\n| **DiffusionBee** | Local | Good | Fast | Local unlimited | Easy setup, open-source |\n| **ComfyUI** | Local | Good | Fast | Local unlimited | Advanced, node-based |\n\n### Free Image Model APIs\n\n| Provider | Model | Free Tier | Notes |\n|----------|-------|-----------|-------|\n| **Replicate** | FLUX.1-schnell | Free tier | Fast inference |\n| **Pollinations** | Various | Unlimited | No signup |\n| **HuggingFace** | SDXL\u002FFLUX | $0.10 credits | Inference API |\n| **Leonardo** | Phoenix | 150 tokens\u002Fday | Commercial OK |\n\n---\n\n## 🎬 Video Generation APIs\n\nText-to-video and image-to-video generation. Hot area in 2026.\n\n| Model | Provider | Quality | Duration | Free Tier | Best For |\n|-------|----------|---------|----------|-----------|----------|\n| **Veo 3** | Google | 🏆 Excellent | 1080p, **60s clips** | Limited preview | Cinematic, realistic |\n| **Sora 3** | OpenAI | 🏆 Excellent | **120s** | ChatGPT Plus | High quality, physics |\n| **Runway Gen-3** | Runway | Excellent | 10 seconds | 3 free credits | Creative, filmmaking |\n| **Pika 3.0** | Pika | Very Good | 3-5 seconds | Free tier | Lip-sync improved |\n| **Luma Dream Machine** | Luma | Very Good | 5 seconds | 30 generations\u002Fmo | Fast, realistic |\n| **Kling** | Kuaishou | Excellent | 2-10 minutes | Limited | Long-form, Chinese |\n| **Hailuo AI** | MiniMax | Good | 6 seconds | Free tier | Character consistency |\n| **Stable Video Diffusion** | Stability | Good | 4 seconds | Local | Open, flexible |\n\n### Video API Pricing (approximate)\n\n| Provider | Cost per video | Generation time |\n|----------|----------------|-----------------|\n| **Runway** | ~$0.20-0.50 | 1-5 min |\n| **Pika** | ~$0.10-0.30 | 30s-2 min |\n| **Luma** | ~$0.30-0.60 | 2-5 min |\n| **Kling** | ~$0.05-0.20 | 1-10 min |\n\n---\n\n## 🌐 AI Browser Automation\n\nTools for AI agents to control browsers - web scraping, form filling, testing.\n\n| Tool | Type | Pricing | Best For |\n|------|------|---------|----------|\n| **Browserbase** | Managed browsers | $5 free tier | Production agents |\n| **Steel.dev** | Browser API | Free tier | AI-native browser control |\n| **Stagehand** | AI browser framework | Open source | Next-gen Playwright |\n| **Playwright** | Browser automation | Free | Reliable, well-documented |\n| **Puppeteer** | Chrome automation | Free | Chrome-specific |\n| **Selenium** | Cross-browser | Free | Legacy support |\n| **Scrapy** | Web scraping | Free | Data extraction |\n\n### AI-Native Browser Tools\n\n| Tool | AI Integration | Use Case |\n|------|----------------|----------|\n| **Stagehand** | Natural language commands | AI agents controlling browsers |\n| **Browserbase** | Session recording for AI | Training agent trajectories |\n| **Steel.dev** | Built for LLM agents | Agent-native browser API |\n\n**Stack Recommendation:**\n- **AI agents:** Stagehand + Browserbase\n- **Web scraping:** Playwright + Scrapy\n- **Testing:** Playwright + AI assertions\n\n---\n\n## 💾 Cheap Vector DB Hosting\n\nProduction-ready vector storage without high costs.\n\n| Provider | Type | Free Tier | Paid | Best For |\n|----------|------|-----------|------|----------|\n| **Supabase Vector** | Postgres + pgvector | 500MB | $25\u002Fmo starter | Full-stack apps |\n| **Neon** | Serverless Postgres | 500MB | $19\u002Fmo | Serverless, branching |\n| **Railway** | Managed Postgres | $5 credits | Usage-based | Easy deployment |\n| **PlanetScale** | MySQL + vectors | 5GB | $39\u002Fmo | Scale, branching |\n| **Chroma Cloud** | Vector-native | Free tier | Usage-based | Pure vector workloads |\n| **Qdrant Cloud** | Vector DB | 1GB | $25\u002Fmo | High performance |\n| **Pinecone** | Managed vector | 2GB | $70\u002Fmo | Production, no ops |\n| **Weaviate Cloud** | Vector DB | 5M vectors | $25\u002Fmo | Hybrid search |\n| **LanceDB** | Embedded\u002FCloud | Free | Cloud beta | Multimodal |\n\n### Self-Hosted (Free Forever)\n\n| Database | Best For | Notes |\n|----------|----------|-------|\n| **ChromaDB** | Prototyping | Simple, Python-native |\n| **Qdrant** | Production | Rust-based, fast |\n| **Milvus** | Enterprise | Scalable, complex |\n| **pgvector** | Postgres apps | Just add extension |\n| **LanceDB** | Embedded | No server needed |\n\n**Recommendation by Stage:**\n- **MVP:** ChromaDB (local) → Supabase (hosted)\n- **Production:** Qdrant Cloud or Pinecone\n- **Enterprise:** Milvus or Weaviate\n\n---\n\n## 🏛️ Common AI Architecture Patterns\n\nProven patterns for building AI applications.\n\n### 1. 🤖 Chatbot Architecture\n\n```\nUser → Chat UI → LLM API → Response\n            ↓\n        Context Memory (Redis\u002FPostgres)\n```\n\n**Stack:**\n- Frontend: Next.js + Vercel AI SDK\n- Backend: FastAPI + OpenRouter\n- Memory: Upstash Redis or Supabase\n\n---\n\n### 2. 📚 RAG Architecture (Like ExamAi)\n\n```\nDocuments → Chunking → Embeddings → Vector DB\n                                    ↓\nUser Query → Embedding → Similarity Search → LLM → Response\n```\n\n**Stack:**\n- Framework: LlamaIndex or LangChain\n- Embeddings: BGE-Large or Jina v3\n- Vector DB: ChromaDB (dev) → Pinecone (prod)\n- LLM: Claude Sonnet [verify] or GPT-4o\n\n---\n\n### 3. 🎯 Agent Architecture\n\n```\nUser Request → Agent Controller → Tool 1 (Search)\n                              → Tool 2 (Code exec)\n                              → Tool 3 (API call)\n                              ↓\n                        Synthesize → Response\n```\n\n**Stack:**\n- Framework: LangGraph, AutoGen, or CrewAI\n- Tools: Function calling with Claude\u002FGPT-4\n- Memory: Vector DB + State management\n- Monitoring: LangSmith or Arize\n\n---\n\n### 4. 🔄 Multi-Model Routing Architecture\n\n```\nUser Request → Router (classify intent)\n                    ↓\n    ┌───────────────┼───────────────┐\n    ↓               ↓               ↓\nCheap Model    Medium Model    Expensive Model\n(GPT-5 Nano)      (Claude Sonnet [verify]) (Claude Opus [verify])\n    ↓               ↓               ↓\nSimple Q&A    Complex task    Hard reasoning\n```\n\n**Implementation:**\n- Router: Fine-tuned classifier or LLM-based\n- Cost optimization: Route 80% to cheap models\n- Fallback: Escalate if cheap model fails\n\n---\n\n### 5. ⚡ Realtime Streaming Architecture\n\n```\nAudio Input → STT → LLM → TTS → Audio Output\n     ↓           ↓      ↓       ↓\n Deepgram   ","该项目是一个精心整理的免费和低成本AI工具列表，涵盖了LLM API、IDEs、代理和基础设施，旨在帮助开发者构建实际的AI应用程序。核心功能包括提供免费的GPT-5\u002FClaude\u002FGemini API访问、低成本的AI API（每百万token仅需0.10-0.50美元）、编码助手及AI原生IDEs（如Cursor, Trae, Windsurf）等。此外，还提供了RAG栈工具、代理框架、本地模型支持以及生产就绪的堆栈配置。适用于希望在不支付高昂月费的情况下快速开发和部署AI应用的开发者。",2,"2026-06-11 02:43:37","CREATED_QUERY"]