Guide
The AI tools directory gathers 101 AI products and platforms people actually use right now, in nine groups: chat assistants (ChatGPT, Claude, Gemini, DeepSeek), writing & office (rewriting, translation, slides, meeting transcription), image & video generation (Midjourney, Firefly, Kling, Hailuo, ComfyUI), voice & music (ElevenLabs, Suno, Whisper, Fish Audio), coding assistants (Cursor, Copilot, Claude Code, Continue), agents & workflows (Dify, n8n, LangChain, MCP), local run (Ollama, LM Studio, llama.cpp), benchmarks & leaderboards (LMArena, Artificial Analysis, OpenCompass), and APIs & model platforms (OpenAI, Anthropic, DeepSeek, SiliconFlow, OpenRouter).
Updated 2026-09-294 sources10 min read
The AI tools directory gathers 101 AI products and platforms people actually use right now, in nine groups: chat assistants (ChatGPT, Claude, Gemini, DeepSeek), writing & office (rewriting, translation, slides, meeting transcription), image & video generation (Midjourney, Firefly, Kling, Hailuo, ComfyUI), voice & music (ElevenLabs, Suno, Whisper, Fish Audio), coding assistants (Cursor, Copilot, Claude Code, Continue), agents & workflows (Dify, n8n, LangChain, MCP), local run (Ollama, LM Studio, llama.cpp), benchmarks & leaderboards (LMArena, Artificial Analysis, OpenCompass), and APIs & model platforms (OpenAI, Anthropic, DeepSeek, SiliconFlow, OpenRouter).
The first question is rarely "which model is strongest" but "can I use it, does it cost anything, do I need an account, and where does my data go". So every entry marks three attributes: Free (a usable free tier), API (you can wire it into your own program or this site’s tools) and Local / this site (the model runs on your own machine or inside toolun). Three toggles at the top of the page combine them — "Free + API" gives the providers you can plug into this site’s AI chat. Providers are ordered international-first: every listing can be signed up for and used from multiple regions, and products available in only one market are not included.
The site's own AI tools sit in a group of their own at the front. They run in BYOK (Bring Your Own Key) mode: you enter your own API key in the settings, the request goes straight from your browser to the provider you picked, and toolun's servers never touch the content. Cutout, OCR and face blur go further and use purely local models, so the images never leave the browser.
With "Free" and "Has an API" selected, the "APIs & model platforms" group gives:
Wiring DeepSeek into this site's AI Chat settings:
APIs & model platforms
Google AI Studio The official Gemini APIs, with a generous free tier. Free · Sign-in · API
OpenRouter One key for many models; easy price comparison and switching. Free tier · Sign-in · API
Cloudflare Workers AI Edge inference with a daily free allowance, integrated with Workers. Free · Sign-in · API
DeepSeek Platform OpenAI-compatible and cheap; paste straight into this site's AI chat. Free tier · Sign-in · API
SiliconFlow Aggregator of open-model APIs; sign-up credit covers Qwen / DeepSeek / FLUX. Free tier · Sign-in · APIEndpoint https://api.deepseek.com/v1
API Key sk-… (created on the DeepSeek Platform)
Model deepseek-chat or deepseek-reasoner
Remember key off = session memory only; on = saved in local localStorageFour things decide it. Unit of measure: by run (N images or N songs a day), by token or character (APIs and TTS), by credit (common in image and video generation, where different operations cost different amounts) or by duration (transcription). Reset cycle: a daily reset is the friendliest, a monthly one needs planning, and a one-off grant ends when it is used up. Quality cuts: free tiers often hand you low resolution, a watermark, an older model or a long queue — the free outputs of remove.bg and Clipdrop only reach preview resolution. Hidden conditions: a card on file (the OpenAI API free credit was discontinued and now needs a top-up), an invitation, an education email address. The only genuinely unlimited free option is running an open model locally, and the price is your hardware and your electricity. When choosing, look at a leaderboard for quality first, at Artificial Analysis for price, and then test the free tier on a real task.
Three ideas often get mixed up. Reachability: whether a site opens directly from where you are; some services block by IP. Registration: even when reachable, sign-up often verifies a phone number or card, and some products are region-locked. Service area: being able to log in does not guarantee the service works — a few products restrict features by region. This directory only lists services you can register for and use from multiple regions; sign-up requirements and regional limits are noted in each card’s review, and if something will not register, a directly usable alternative usually sits in the same group.
Data can go three ways when you use an AI service. The provider’s own product (ChatGPT web, Claude app): content is stored with the provider and often used for training by default; this can be turned off in settings. Third-party wrappers: content passes through another server before reaching the model provider, adding a leak risk that is rarely disclosed. BYOK tools (this site’s AI chat, Continue, Aider, Open WebUI with a cloud API): your key, your request, straight from your device to the provider with no middleman; history stays local. This site keeps keys in session memory by default and only writes to localStorage when you tick "remember key". Local run (Ollama, LM Studio, Whisper, ComfyUI, this site’s cutout / OCR): data never leaves the device — the only worry-free route for contracts, medical records and unreleased code.
Text models are estimated from parameter count: a 7–9B model at 4-bit quantisation takes about 5 GB of memory and runs on a laptop; 14B takes about 9 GB; 32B takes about 20 GB and needs a discrete GPU or a Mac with 32 GB or more of unified memory; 70B and above is workstation territory. Ollama and LM Studio pick a GGUF quantisation level automatically and split work between CPU and GPU. Image models: SDXL needs 6–8 GB of VRAM, FLUX is comfortable above 12 GB; video models generally want 16–24 GB. Whisper transcription runs in real time on the CPU through whisper.cpp. Unified memory makes Apple silicon very good value for local inference; on Windows, prefer an NVIDIA card, since the CUDA ecosystem is the most complete.
LMArena runs anonymous battles with user votes, which reflects “which answer people prefer” and is most useful for chat tasks — though it favours verbose, prettily formatted answers. LiveBench rotates its questions to fight training-set contamination; SWE-bench and the Aider leaderboard measure coding on real tasks; Artificial Analysis shows quality, speed and price on three axes and is the first place to look when picking an API provider; OpenCompass and SuperCLUE cover Chinese-language tasks. Any single leaderboard can be gamed, so pick a model from the overlap of two or three and then test it on your own real tasks.
Going from "chatting with a model" to "letting a model do things for me" needs three ingredients: tool calling (the model can search, read files, call APIs), multi-step planning (deciding the next step from results), and state plus memory. Dify and Flowise assemble these visually; n8n and Zapier grew AI nodes from the automation side; LangGraph, CrewAI, AutoGen and the OpenAI Agents SDK are code-first frameworks. MCP (Model Context Protocol) solves how tools plug in uniformly: a tool exposed by one MCP server works in Claude, Cursor, ChatGPT and any other MCP-capable client, without writing it again per client.
In the "Chat assistants" group, ChatGPT, Claude, Gemini and DeepSeek all have free tiers; start with those for everyday Q&A and consider a subscription only when you need longer context or stronger reasoning.
Under "APIs & model platforms", tick "Free + Has an API", register with DeepSeek or SiliconFlow and paste the key into AI chat.
The “Local running” group: Ollama + Open WebUI for local question answering, AnythingLLM for a local knowledge base, whisper.cpp to transcribe recordings; for images, use the local engines in this site's AI Cutout and Dual-engine OCR.
Coding assistants split by workflow: in-IDE completion (GitHub Copilot, Tabnine), AI-native editors (Cursor, Windsurf), terminal agents (Claude Code, Aider), whole-site generation (v0, bolt.new, Lovable); if code must stay inside your network, pick Continue + Ollama.
Listings are limited to services you can register for and use from multiple regions; products that need a region-specific account, exist in only one market or have no international version are excluded, which is why a batch of regional-only products is absent. The same capability usually has a substitute among the listed tools.
No. Requests go from your browser straight to the provider's address, and toolun has no relay server; the key lives in session memory only by default and disappears when you close the page, and it is written to local localStorage only when you tick “Remember key”. You can confirm the request target in the network panel of your browser's developer tools.
Midjourney dropped its free trial after 2023 and requires a subscription; a free ChatGPT account can generate images, with a daily cap. The “Free” label asks whether a usable free tier exists at all.
Ollama is a command line plus a local API, good for plugging into other tools (Open WebUI, Continue, this site's AI Chat); LM Studio has a graphical interface for chatting and picking models. Both speak OpenAI-compatible endpoints and can coexist.
Not necessarily. A leaderboard measures general ability, while your task may care more about Chinese, long documents, code or price; look at the overlap between two or three leaderboards and then run your own real examples on the free tier.
The directory data loads statically with the page; search, filtering and bookmarks happen locally in the browser, and this page sends no requests and records no clicks. Outbound links open in a new tab with rel="noopener noreferrer" and are then governed by the other site's privacy policy. The on-site AI tools are BYOK: the key and the conversation travel only between your browser and the provider you chose; the local-model tools (cutout, OCR, face blur) never let your data leave the browser.
Updated 2026-09-29
101 AI tools and platforms for chat, writing, office work, image and video, voice and music, coding, agents, local models, rankings, and APIs, with free, API and local-run badges
Bring your own OpenAI-compatible key (DeepSeek, SiliconFlow, Ollama and more); chat history stays on your machine.
Rewrite, expand, shorten and change tone, with templates so you are not writing the prompt from scratch every time.
Prefers a self-hosted DeepLX endpoint, falls back to an LLM, and translates long texts in segments.
Connects to OpenAI-compatible image endpoints; results are kept in the browser.
Local Tesseract and cloud OCR.space side by side; sensitive files stay local.
The model is downloaded to the browser and then runs offline; images are never uploaded.
A cloud fallback when the local cutout is not good enough; needs your own remove.bg key.
Detects faces locally and blurs them, protecting privacy before you post a photo.
The most mainstream entry point: basic models work without signing in, while Plus unlocks advanced models and enough image generation.
Well regarded for long text and writing; Artifacts is great for building docs and code in conversation. Sign-up needs a phone number from a supported region.
Works with a Google account, integrates with Docs and Gmail, and is strong on long context; unavailable in some regions.
Free access to GPT-series models with web search, right from the Edge sidebar.
AI search with cited sources; more reliable than plain chat for research.
The chat product from Europe's open-model company: fast responses and a generous free tier.
From xAI, with live X platform content and a direct style.
One account to switch between many models; the free allowance is credit-based.
The official chat for an open-weight model family: reasoning mode is free and API pricing is among the lowest.
Summarise, continue and translate inside Notion pages, as a paid add-on.
The veteran English grammar and tone checker; the free tier is enough for everyday email.
Rewrites and polishes English, German, French and Spanish with a natural style; no login needed.
Long the leader in machine translation quality, with a character limit on the free tier; this site's translator can use its API.
One sentence in, an editable slide deck or web page out; free credit system.
Turns a passage of text into a diagram or flowchart; quick visuals for reports.
Real-time transcription and summaries for English meetings.
Upload sources and ask questions grounded only in them; it can also generate a podcast-style audio overview.
The strongest artistic text-to-image model: no free trial, subscription required; the web version no longer needs Discord.
Draw and edit images directly inside ChatGPT, with accurate text rendering; free with a quota.
Cleanly licensed training data, the safest bet for commercial work; free credit system.
The most accurate text rendering among text-to-image models; great for poster headlines.
Plenty of game and concept-art styles, with a daily free allowance.
A pioneer in video generation and editing, with the Gen series; few free credits.
Short-video effects and image-to-video, with plenty of social-friendly tricks.
Stable image-to-video quality, with a limited number of free generations per month.
The classic local text-to-image UI with the largest plug-in ecosystem; needs 6 GB+ of VRAM.
A node-based local generation tool that runs FLUX and video models; there is a learning curve.
Local text-to-image that hides all the SDXL parameters and just takes a prompt, like Midjourney.
Real-time generation, upscaling and restoration, with a daily free allowance.
Single-purpose tools for cutout, object removal and relighting; the free tier limits resolution, and cutout is available locally on this site.
This site: local browser models for photo restoration and colourisation.
A leading video generator with good motion range and physicality; daily free credits.
Video generation and voice cloning, with a daily free allowance.
The quality benchmark for speech synthesis and cloning; 10,000 free characters per month.
One sentence in, a complete song with lyrics out; 10 free songs a day.
The other leader in music generation, with more natural vocals and continuation.
Open-source speech recognition, accurate across languages, and completely free when run locally.
Whisper's C++ port runs on laptop CPUs, with a one-click app on Mac.
A desktop GUI for Whisper: import audio or video and get subtitles.
Turns a phone recording into studio-quality audio in one click; free for a limited time each day.
TTS through Edge's online speech endpoint; the command line is free and there are many Chinese voices.
This site: ffmpeg-based noise reduction, nothing uploaded.
An open-model service with good Chinese voice cloning.
An AI-native editor whose agent mode edits across files; the free tier limits requests.
Completion and chat inside VS Code and JetBrains; the personal free tier has a monthly completion limit.
Agent-style coding in the terminal: reads the whole repo, changes code and runs tests, billed by API usage.
Codeium's AI IDE with a fairly generous free tier.
An open-source IDE AI plug-in that can use a local Ollama model, so code never leaves your network.
Open-source command-line AI coding with good Git integration and support for any model API.
Describe an interface and get React + Tailwind components; extremely fast for frontend prototypes.
Generate and run a full-stack project in the browser; free tokens are limited.
Chat your way to a web app with Supabase integration; prototyping for non-programmers.
An agent inside a cloud IDE, from requirements to deployment.
A completion tool focused on privacy and private deployment, aimed at enterprises.
Open-source automation workflows (a Zapier alternative) with a full set of AI nodes; self-hostable.
The most popular LLM application framework; LangGraph builds stateful agents.
OpenAI's lightweight official agent framework, with handoffs and guardrails built in.
A multi-agent collaboration framework that assigns work by role.
Microsoft's multi-agent conversation framework.
A drag-and-drop LangChain flow editor, open source.
The SaaS automation veteran, with plenty of AI steps; the free tier allows 100 tasks a month.
Visual automation that is more flexible and cheaper than Zapier.
An open protocol that lets models call external tools and data; already supported by Claude, Cursor, ChatGPT and others.
An open-source LLM app platform for visually orchestrating RAG and agents; self-hostable.
An open-source knowledge-base Q&A platform that can be self-hosted.
One command pulls and runs open models (Llama, Qwen, DeepSeek) and exposes an OpenAI-compatible API.
A GUI local model runner: search, download, chat and start a local server.
The low-level engine for local inference and the origin of the GGUF format; runs on CPU or GPU.
An open-source local ChatGPT alternative with a friendly interface.
Local chat and local document Q&A with low hardware requirements.
A self-hosted web UI for Ollama, with multi-user support, RAG and plug-ins.
A desktop local knowledge-base Q&A app: drop in a document and ask.
The central hub for open models, where weights and model cards live.
The AUTOMATIC1111 of local LLMs, with every parameter exposed.
Anonymous battles ranked by user votes; the leaderboard closest to real experience.
Compare models and API providers across quality, speed and price.
A benchmark whose questions are refreshed regularly to resist gaming.
A leaderboard that evaluates coding agents on real GitHub issues.
Hugging Face's standardised benchmark for open models.
Tests models on real code-editing tasks, as a reference for coding choices.
Aggregates officially published benchmark scores and prices for a one-screen comparison.
An open evaluation system and leaderboard from Shanghai AI Lab, with broad Chinese-task coverage.
A Chinese LLM benchmark split by capability dimension.
The GPT API family and the de facto standard for OpenAI-compatible formats.
The Claude API; sign-up requires a supported region.
A generous free tier for the Gemini API, with parameters you can tweak right on the page.
One key for hundreds of models in a unified OpenAI format, with free models available.
Inference on custom silicon makes open models extremely fast; a free tier is available.
A hosting API for open models, with sign-up credit.
Run open image, video and audio models by the second with a one-line API.
Open models on edge nodes, with a free daily allowance and Workers integration.
OpenAI-compatible and extremely cheap; can be pasted straight into this site's AI chat.
An open-model API platform with sign-up credit; Qwen, DeepSeek and FLUX are all available.