AI resource hub

Pick the right AI tool, then actually ship with it

Comparisons and hands-on guides for AI tools, LLMs, coding agents and API gateways — including the AI coming out of China — plus prompt templates and an AI wiki.

Common searchesAI WritingAI ImagesAI VideoBackground RemovalPPTCoding Assistants

Tools

AI Tools

Start with popular or recently added tools, then review features, pricing, fit, and alternatives.
DeepSeekChat

DeepSeek's LLM products: V4-Pro and V4-Flash both ship a native 1M-token context window, model weights are open-sourced, and the web app and mobile apps are free to use.

  • Free tier · Partial
Nano BananaImage Generation

Google's image model: Nano Banana 2 (Gemini 3.1 Flash Image) is now the default generator across Gemini and Search, alongside Pro and Lite variants.

  • Free tier · Partial
Tongyi QianwenChat

Alibaba's Qianwen assistant — the web entry moved from tongyi.com to qianwen.com; the current flagship model is Qwen3.8-Max and the open-weight releases continue

  • Free tier · Partial
  • Chinese support · Yes
  • Works in China · Yes
Wenxin YiyanChat

Baidu's ERNIE chat entry point — the web app remains at yiyan.baidu.com and the mobile app is now called Wenxiaoyan; it currently runs ERNIE 5.0 / 5.1 with deep Baidu Search integration

  • Free tier · Partial
  • Chinese support · Yes
  • Works in China · Yes
Stable DiffusionImage Generation

Stability AI's open-source text-to-image model from 2022: public weights, local deployment and a large ecosystem of fine-tunes.

Flux.1Image Generation

Black Forest Labs' text-to-image model family; the current FLUX.2 includes API-only pro, flex, and max plus open-weight dev and klein, with licenses that vary by model

ChatGPTChat

OpenAI's AI chat product for Q&A, writing, code, data analysis, image generation, and web search; it moved fully to GPT-6 in October 2026, and the Free and Go tiers show ads

  • Free tier · Partial
Notion AIWriting

Notion's built-in AI: generate, rewrite, summarize and translate inside any doc, and answer questions from your own Notion content.

  • Free tier · Partial
Metaso Writing CatWriting

Metaso's Chinese writing assistant — real-time correction, rewriting and multi-user collaboration across web, mini program, a Word add-in and browser extensions

  • Free tier · Partial
  • Chinese support · Yes
  • Works in China · Yes
Niji JourneyImage Generation

The anime image generator from Midjourney and Spellbrush; Niji 7 is tuned for Japanese-style characters, illustrations and scenes.

AnthropicChat

The company behind Claude: model documentation, API references and enterprise entry points (for the chat product, see Claude).

  • Free tier · Partial
  • Works in China · No
Duiyou AI Creation ToolPopular Picks

A free AI illustration and design platform from Alibaba Design, with a community for sharing work and prompts.

  • Free tier · Yes
  • Chinese support · Yes
  • Works in China · Yes
Browse all AI toolsView trending rankings

Comparisons

Selected comparisons

Two products at a time, compared row by row, with a clear verdict on which to pick.
View all comparisons

Prompts

New prompts you can use now

Each entry includes a concrete task and a complete prompt you can adapt immediately.
Image GalleryOutfits
Winter outfits · 8-panel grid

An outfit-grid prompt inferred panel by panel by Codex from a verified Xiaohongshu reference image, with a newly generated complete collage.

Open prompt →
Image GalleryOutfits
University outfits · 9-panel grid

An outfit-grid prompt inferred panel by panel by Codex from a verified Xiaohongshu reference image, with a newly generated complete collage.

Open prompt →
Image GalleryOutfits
Travel outfits · 9-panel grid

An outfit-grid prompt inferred panel by panel by Codex from a verified Xiaohongshu reference image, with a newly generated complete collage.

Open prompt →
Image GalleryPhotography
Seaside travel photos: six natural poses

Six seaside poses inferred from the first six panels of a Xiaohongshu reference image and regenerated in two columns and three rows. The face is newly designed and age 25 is a generation constraint. This is not the original creator’s prompt or photograph.

Open prompt →
Image GalleryOutfits
Teacher workwear · 9-panel grid

An outfit-grid prompt inferred panel by panel by Codex from a verified Xiaohongshu reference image, with a newly generated complete collage.

Open prompt →
AI Women PortraitsOutfits
Colorful feminine outfits · Nine-panel grid

An outfit-grid prompt inferred by Codex from a Xiaohongshu outfit collage, with a newly generated complete image showing one person in multiple looks.

Open prompt →
Browse all prompts

Learn

AI Tutorials

Use practical steps, constraints, and alternatives to make better tool choices.
Hybrid AI coding: an acceptance guide for API spend, time and output qualityWhat Is Jev? An AI Model That Can't Chat, and How to Split Work With LLMsCan local models replace an AI subscription? Capability, performance and cost before buying a Mac StudioSet Permissions Before You Let an AI Agent Use Your Computer: A Least-Privilege Checklist for Mac, Windows and Coding AgentsInstalling and Using 9Router: One Local Endpoint for Claude Code, Codex and Cursor

Wiki

AI Wiki quick reference

Concise definitions for understanding AI concepts before choosing a tool.
Agent SandboxAn agent sandbox is an isolated environment in which an AI agent runs code, reads and writes files, and uses the network, keeping what it can reach within controlled boundaries. Several 2026 incidents in which agents got out of their sandboxes showed that inconspicuous channels such as DNS, caches, and config files are boundaries too.AI WatermarkingAI watermarking embeds a machine-detectable mark in AI-generated text, images, audio, or video to show that the content was made by AI. Article 50 of the EU AI Act has applied since August 2, 2026, and mainstream models such as Claude began watermarking by default.Reward HackingReward hacking is when a model exploits loopholes in the scoring rules to earn a high score without doing what it was actually meant to do. In 2026 it moved from a technical training problem to a real-world risk that inflates benchmark scores and can even push agents past their boundaries.SandbaggingSandbagging is when an AI model deliberately performs worse on a capability evaluation than it really can. Safety decisions increasingly rest on evaluation results, and if a model can recognize a test and underperform on purpose, those results lose their footing. In 2026, the GPT-6 Astra system card brought the issue into the open.Third-party EvaluationThird-party AI evaluation means independent organizations outside the model developer assessing an AI model's capabilities and risks. In September 2026, Anthropic and OpenAI both committed to giving independent evaluators employee-like access, the AEF-1 standard began drawing lab sign-ons, and the debate shifted to who pays and who can publish the conclusions.Agent MemoryAgent memory is the machinery that lets an agent persist information outside the context window and retrieve it on demand. The problem it solves isn't "the window is too small" but "remember the right things across sessions, days, and tasks — and update them when facts change"

GitHub AI Weekly

What is gaining on GitHub this week?

Preview the latest trend call and the fastest-rising open-source projects.
GitHub AI Weekly · Issue 10 (Oct 17, 2026) · AI Beyond

More important than keeping an agent running is letting people inspect evidence, revise intermediate results, and recover compressed details when needed.

  1. 01
    morluto/reaAI agents
    +248,962 This week
  2. 02
    storytold/artcraftMultimodal
    +20,267 This week
  3. 03
    mattpocock/skillsAI coding
    +11,937 This week
  4. 04
    cathrynlavery/diagram-designAI coding
    +9,024 This week
  5. 05
    heygen-com/hyperframesMultimodal
    +5,601 This week
Read this issueView weekly archive

Gateways

API Gateways

Compare gateway providers, status, cost, and pre-launch checks in one place.
Compare providersService statusPricing calculatorPre-launch checks

News

AI News

Connect model, product, and risk updates to your own tool decisions.
Anthropic Puts 5% of Its Shares Into a Public Benefit Fund: Another 2% for Every 50% Rise in Value, None for Lobbying or Political GivingGoogle Lets AI Agents Run Their Own Wet-Lab Experiments: Rehearsing in a Digital Twin of the Bench, Aptamer Hit Rates Rose From 0 to 42.1% in Three RoundsThe Lean Proof Assistant Now Ships With Four Extra Kernels: con-leche Was Written and Proven Consistent by Claude, and lake check --paranoid Rechecks With All of ThemChatGPT Slides Starts Rolling Out: Teams Edit Decks Together in Space, With Web Export to PowerPoint, PDF and Google SlidesClaude Traced the NIST Elliptic Curve Seeds in Under Half an Hour: A 1997 Clock Reading and a C Library Random Number Function, With the $36,864 Bounty Going to ISRG