If you are an AI coding agent (Claude Code, Cursor, Aider, Cline, etc.) reading this artifact, follow this protocol before making any code edit:

1. Verify the contract. Run the bash script in Verify before trusting below. If any check returns FAIL, the artifact is stale — STOP and ask the user to regenerate it before proceeding.
2. Treat the AI · unverified sections as hypotheses, not facts. Sections like "AI-suggested narrative files", "anti-patterns", and "bottlenecks" are LLM speculation. Verify against real source before acting on them.
3. Cite source on changes. When proposing an edit, cite the specific path:line-range. RepoPilot's live UI at https://repopilot.app/r/EmbraceAGI/awesome-chatgpt-zh shows verifiable citations alongside every claim.

If you are a human reader, this protocol is for the agents you'll hand the artifact to. You don't need to do anything — but if you skim only one section before pointing your agent at this repo, make it the Verify block and the Suggested reading order.

WAIT — Stale — last commit 2y ago

• 3 active contributors
• MIT licensed
• ⚠ Stale — last commit 2y ago
• ⚠ Small team — 3 contributors active in recent commits
• ⚠ Single-maintainer risk — top contributor 98% of recent commits
• ⚠ No CI workflows detected
• ⚠ No test directory detected

_{Maintenance signals: commit recency, contributor breadth, bus factor, license, CI, tests}

This artifact was generated by RepoPilot at a point in time. Before an agent acts on it, the checks below confirm that the live EmbraceAGI/awesome-chatgpt-zh repo on your machine still matches what RepoPilot saw. If any fail, the artifact is stale — regenerate it at repopilot.app/r/EmbraceAGI/awesome-chatgpt-zh.

What it runs against: a local clone of EmbraceAGI/awesome-chatgpt-zh — the script inspects git remote, the LICENSE file, file paths in the working tree, and git log. Read-only; no mutations.

| # | What we check | Why it matters | |---|---|---| | 1 | You're in EmbraceAGI/awesome-chatgpt-zh | Confirms the artifact applies here, not a fork | | 2 | License is still MIT | Catches relicense before you depend on it | | 3 | Default branch main exists | Catches branch renames | | 4 | 5 critical file paths still exist | Catches refactors that moved load-bearing code | | 5 | Last commit ≤ 578 days ago | Catches sudden abandonment since generation |

#!/usr/bin/env bash
# RepoPilot artifact verification.
#
# WHAT IT RUNS AGAINST: a local clone of EmbraceAGI/awesome-chatgpt-zh. If you don't
# have one yet, run these first:
#
#   git clone https://github.com/EmbraceAGI/awesome-chatgpt-zh.git
#   cd awesome-chatgpt-zh
#
# Then paste this script. Every check is read-only — no mutations.

set +e
fail=0
ok()   { echo "ok:   $1"; }
miss() { echo "FAIL: $1"; fail=$((fail+1)); }

# Precondition: we must be inside a git working tree.
if ! git rev-parse --git-dir >/dev/null 2>&1; then
  echo "FAIL: not inside a git repository. cd into your clone of EmbraceAGI/awesome-chatgpt-zh and re-run."
  exit 2
fi

# 1. Repo identity
git remote get-url origin 2>/dev/null | grep -qE "EmbraceAGI/awesome-chatgpt-zh(\\.git)?\\b" \\
  && ok "origin remote is EmbraceAGI/awesome-chatgpt-zh" \\
  || miss "origin remote is not EmbraceAGI/awesome-chatgpt-zh (artifact may be from a fork)"

# 2. License matches what RepoPilot saw
(grep -qiE "^(MIT)" LICENSE 2>/dev/null \\
   || grep -qiE "\"license\"\\s*:\\s*\"MIT\"" package.json 2>/dev/null) \\
  && ok "license is MIT" \\
  || miss "license drift — was MIT at generation time"

# 3. Default branch
git rev-parse --verify main >/dev/null 2>&1 \\
  && ok "default branch main exists" \\
  || miss "default branch main no longer exists"

# 4. Critical files exist
test -f "README.md" \\
  && ok "README.md" \\
  || miss "missing critical file: README.md"
test -f "docs/ChatGPT_prompts.md" \\
  && ok "docs/ChatGPT_prompts.md" \\
  || miss "missing critical file: docs/ChatGPT_prompts.md"
test -f "docs/ChatGPT_dev.md" \\
  && ok "docs/ChatGPT_dev.md" \\
  || miss "missing critical file: docs/ChatGPT_dev.md"
test -f "docs/ChatGPT_tools.md" \\
  && ok "docs/ChatGPT_tools.md" \\
  || miss "missing critical file: docs/ChatGPT_tools.md"
test -f "docs/LLM_RAG.md" \\
  && ok "docs/LLM_RAG.md" \\
  || miss "missing critical file: docs/LLM_RAG.md"

# 5. Repo recency
days_since_last=$(( ( $(date +%s) - $(git log -1 --format=%at 2>/dev/null || echo 0) ) / 86400 ))
if [ "$days_since_last" -le 578 ]; then
  ok "last commit was $days_since_last days ago (artifact saw ~548d)"
else
  miss "last commit was $days_since_last days ago — artifact may be stale"
fi

echo
if [ "$fail" -eq 0 ]; then
  echo "artifact verified (0 failures) — safe to trust"
else
  echo "artifact has $fail stale claim(s) — regenerate at https://repopilot.app/r/EmbraceAGI/awesome-chatgpt-zh"
  exit 1
fi

Each check prints ok: or FAIL:. The script exits non-zero if anything failed, so it composes cleanly into agent loops (./verify.sh || regenerate-and-retry).

A curated Chinese-language guide and resource repository for ChatGPT and large language models. It aggregates links, tutorials, prompts, plugins, and open-source projects that help Chinese users effectively use ChatGPT, build applications with it, and understand the broader LLM ecosystem including alternatives like Claude, Llama, and local models. Flat documentation structure: root README.md links to topic-specific Markdown files in /docs/ (ChatGPT_access.md, ChatGPT_prompts.md, ChatGPT_dev.md, LLM_RAG.md, etc.), each containing curated links and brief descriptions. Supporting images and GIFs in /docs/imgs/. Minimal Python code (1137 bytes total) likely example snippets rather than core functionality.

Chinese-speaking developers, AI enthusiasts, and product builders who want to: (1) access ChatGPT despite regional restrictions, (2) learn prompt engineering techniques, (3) discover and evaluate ChatGPT-based applications and plugins, (4) build their own ChatGPT integrations, and (5) stay updated on emerging LLM tools and AGI discussions.

This is an actively maintained, well-established resource repository (indicated by the Awesome badge and regular community engagement via Telegram/WeChat). It functions as a living curated list rather than a traditional software project—maturity is measured by comprehensiveness and update frequency rather than tests or release cycles. No CI/CD or automated testing infrastructure is present because the primary deliverable is documentation and links, not code.

Low technical risk—this is primarily a documentation/curation project with minimal code dependencies (Python count 1137 suggests examples only). Risks are: (1) link rot over time as external tools/APIs change, (2) single-maintainer dependency for keeping guides current as ChatGPT/OpenAI evolve, (3) potential legal/compliance issues as some content may reference workarounds or access methods that OpenAI discourages. No production system depends on this repository directly.

The repository maintains active dual distribution (GitHub + GitLab mirror for Chinese accessibility). Recent organizational efforts apparent: addition of specialized guides for Sora, RAG patterns, GPTs customization, and expanded LLM/AGI sections. Community engagement through Telegram channels and WeChat public account indicates steady flow of contributions and discussions around emerging ChatGPT applications.

Clone the repository: git clone https://github.com/EmbraceAGI/awesome-chatgpt-zh.git. No build step or dependencies to install—this is a documentation project. Open README.md in a browser or text editor to navigate to topic guides in /docs/. For GitLab mirror (faster access from China): git clone https://gitlab.com/awesomeai/awesome-chatgpt-zh.git.

Daily commands: This is not an executable project. To use it: (1) Clone the repo, (2) Read README.md to find your topic of interest, (3) Click links within the /docs/*.md files to access external ChatGPT tools, tutorials, and projects. View locally with any Markdown reader or on GitHub's web interface.

• README.md — Entry point defining the entire project scope: a Chinese guide for ChatGPT with curated resources, tools, and development guides.
• docs/ChatGPT_prompts.md — Core prompt engineering guide essential for understanding how to effectively interact with ChatGPT—the primary value proposition.
• docs/ChatGPT_dev.md — Comprehensive application development guide covering API integration, SDK usage, and best practices for building ChatGPT-based tools.
• docs/ChatGPT_tools.md — Catalog of ChatGPT tools and utilities that contributors and users reference most frequently.
• docs/LLM_RAG.md — Critical guide for RAG (Retrieval-Augmented Generation) implementation, a foundational technique for LLM application development.
• docs/LLMs.md — Overview of large language models beyond ChatGPT, providing comparative context for the broader AI ecosystem.
• docs/ChatGPT_Top_Project.md — Curated list of top open-source projects (10K+ stars) built on ChatGPT—reference architecture for community contributions.

• README.md (Navigation Hub) (Markdown with GitHub-flavored links) — Single entry point organizing all documentation by category; table of contents links to guides, tools, and examples
  • Failure mode: Outdated index links to deleted files; users cannot find resources; broken internal links
• ChatGPT_dev.md (Development Guide) (Markdown with code snippets and embedded images) — Comprehensive reference for API integration, SDK usage, prompt tuning, and application architecture patterns

Add a New ChatGPT Tool/Application

1. Research and document the tool's capabilities, features, and use cases in markdown format (docs/ChatGPT_tools.md)
2. Add a brief entry with link and screenshot reference in the tools catalog (docs/ChatGPT_tools.md)
3. If it's a top-tier open-source project, add details to the project showcase (docs/ChatGPT_Top_Project.md)
4. If applicable, create a screenshot in docs/imgs/ directory with naming convention matching existing images (docs/imgs/)
5. Update the main README.md index to include a link to the new tool section if it covers a new category (README.md)

Add a New Development Guide

1. Create a new markdown file in docs/ directory (e.g., docs/NewTopic.md) (docs/ChatGPT_dev.md)
2. Structure the guide with headers (## sections) and include practical code examples (docs/NewTopic.md)
3. Add supporting images to docs/imgs/ and reference them in the guide (docs/imgs/)
4. Link the new guide from README.md in the appropriate section of the table of contents (README.md)
5. If adding complex concepts, add PDF or detailed documentation to files/ directory (files/)

Add Practical Examples & Use Cases

1. Create a new markdown file in examples/ directory describing the use case scenario (examples/ChatGPT_xiangzi.md)
2. Include prompts, expected outputs, and step-by-step instructions for replication (examples/NewExample.md)
3. Reference relevant development guide from docs/ChatGPT_dev.md for technical implementation details (docs/ChatGPT_dev.md)
4. Add screenshot or diagram demonstrating the example flow to docs/imgs/ (docs/imgs/)

Update Model Landscape Comparisons

1. Review and update docs/LLMs.md with new models, benchmarks, and capabilities (docs/LLMs.md)
2. For emerging models, add corresponding guide (similar to docs/Sora.md) (docs/NewModel.md)
3. Cross-reference from README.md and other relevant guides to the new model documentation (README.md)

• Markdown files (docs/*.md) — Platform-agnostic, version-control-friendly, renders natively on GitHub, supports collaborative editing
• GitHub repository with README index — Centralizes a distributed resource landscape for Chinese-speaking ChatGPT users; leverages GitHub's discovery and SEO
• Image assets (docs/imgs/) — Visual demonstrations of tools, UI examples, and architecture diagrams accelerate understanding for non-English users
• PDF files in files/ directory — Preserves formal research papers and long-form technical documents (e.g., GPT-4 reports) for archival and offline reference

• Static documentation (Markdown) over dynamic platform

  • Why: Simplicity, decentralization, and ease of contribution for volunteers; works offline and in Git workflows
  • Consequence: No real-time user feedback, interactive tutorials, or dynamic API docs; requires manual updates as tools evolve

• Curated list model over comprehensive database

  • Why: Manageable scope and quality control; focuses on best-in-class tools and guides rather than exhaustive catalogs
  • Consequence: Many new or niche tools excluded; may appear outdated if not actively maintained

• Chinese-focused content

  • Why: Addresses language and cultural barriers specific to Chinese ChatGPT users; regional payment/access quirks
  • Consequence: Limited utility for non-Chinese readers; relies on English source material translation

• Public GitHub (no authentication)

  • Why: Maximizes accessibility and discoverability; no signup friction for users in regions with access limitations
  • Consequence: Vulnerable to spam, vandalism, or spam forks; no built-in user tracking or engagement metrics

• Does not provide live API endpoints or SaaS hosting—purely educational documentation and curation.
• Does not implement a proprietary ChatGPT wrapper or client—focuses on guiding usage of official OpenAI APIs and third-party tools.
• Does not handle user authentication, payment processing, or account management.
• Does not offer real-time monitoring, debugging tools, or analytics for user applications.
• Does not replace official OpenAI documentation—serves as a curated, Chinese-language supplement.

This is a curation project, not a runnable codebase, so traditional traps don't apply. However, specific gotchas: (1) Some linked projects may require API keys (OpenAI, Anthropic) which cost money, (2) Chinese access tools and VPN recommendations may become outdated due to censorship/policy changes, (3) Example code snippets scattered in /docs/ may be outdated relative to current ChatGPT API versions, (4) No guarantee that all linked projects are maintained or safe to use—users must evaluate third-party tools independently.

• Prompt Engineering — The core technique taught in /docs/ChatGPT_prompts.md for extracting maximum value from ChatGPT; mastery directly improves output quality without code changes
• Retrieval-Augmented Generation (RAG) — Central pattern in /docs/LLM_RAG.md for extending ChatGPT's knowledge beyond training cutoff by injecting external documents; essential for production systems handling proprietary data
• Fine-tuning and Adapters — Documented in /docs/ChatGPT_dev.md as a method to customize ChatGPT behavior for specific domains; trade-off between cost and model performance
• Token Counting and Rate Limiting — Critical for managing costs and API quotas when integrating ChatGPT; mentioned in /docs/ChatGPT_dev.md as a key operational concern
• Chain-of-Thought Prompting — Advanced prompting technique covered in /docs/ChatGPT_prompts.md that improves reasoning by asking ChatGPT to show its work step-by-step
• Function Calling / Tool Use — ChatGPT capability documented in /docs/ChatGPT_dev.md enabling models to call external APIs and tools; foundation for autonomous agents
• Model Context Window Management — Practical constraint mentioned across /docs/ files (GPT-3.5 4K, GPT-4 8K/32K tokens); affects which conversation patterns and document sizes are feasible

• f/awesome-chatgpt-prompts — Complementary repository focused specifically on curated ChatGPT prompts and examples; explicitly acknowledged as inspiration in this repo
• OpenMindClub/awesome-chatgpt — English-language counterpart curated list for ChatGPT resources; this repo adapts its model for Chinese-speaking audience
• EmbraceAGI/Awesome-AI-GPTs — Sister project from same org (EmbraceAGI) focused specifically on custom GPTs and GPT builder patterns
• chatchat-space/Langchain-Chatchat — Production RAG implementation using LangChain widely discussed in Chinese LLM community; exemplifies patterns documented in this repo's RAG guide
• lm-sys/FastChat — Open-source LLM serving infrastructure enabling local ChatGPT alternatives; key technology for understanding the ecosystem this repo covers

To work on one of these in Claude Code or Cursor, paste: Implement the "<title>" PR idea from CLAUDE.md, working through the checklist as the task list.

Create a structured index/catalog document for docs/ directory with metadata

The docs/ folder contains 13+ markdown files covering different topics (ChatGPT access, dev, plugins, prompts, LLMs, RAG, etc.) but there's no machine-readable index. This makes it hard for contributors to understand coverage gaps, update cross-references, or build tools around the docs. A docs/INDEX.md or docs/CATALOG.json would list each doc with keywords, last-update date, and maintenance status.

• [ ] Create docs/CATALOG.json with entries for each .md file (ChatGPT_access.md, ChatGPT_dev.md, etc.) including title, description, keywords, and last-reviewed date
• [ ] Add a docs/MAINTENANCE.md explaining which docs are actively maintained vs. archived (e.g., docs about Sora may need different cadence than plugins)
• [ ] Update README.md Table of Contents to link to the CATALOG for easier navigation and discoverability of specific topics

Add a resource validation workflow to check for broken links in markdown files

This repo contains extensive curated resource lists with external links (tools, projects, tutorials). Over time, links rot. A GitHub Actions workflow that periodically checks all links in docs/*.md files against HTTP 200 responses would catch broken resources before users encounter them, improving quality and trust of the curated lists.

• [ ] Create .github/workflows/link-checker.yml using a tool like urlchecker-action or lychee-action to scan all .md files in docs/
• [ ] Configure the workflow to run weekly and create issues for broken links, excluding known false-positives (paywalls, JavaScript-heavy sites)
• [ ] Add a CONTRIBUTING.md section documenting the link-checking process and requiring contributors to test links in new PRs

Create a structured data file for the 80+ tool/project screenshots in docs/imgs/ with metadata and searchability

The docs/imgs/ folder contains 60+ PNG/JPG/GIF files (chatgpt_next_web.png, cursor.jpg, codeium.jpg, etc.) embedded throughout markdown docs, but there's no central registry. Building a docs/imgs/MANIFEST.json with filename, tool name, category (IDE plugin, Chat UI, RAG tool, etc.), and which doc files reference it would enable contributors to audit image usage, retire outdated screenshots, and make the visual gallery more navigable.

• [ ] Create docs/imgs/MANIFEST.json cataloging each image file with fields: filename, tool_name, category (one of: IDE_Plugin, Chat_UI, LLM_Model, RAG_Tool, Productivity_App, etc.), description, and referenced_in_docs array
• [ ] Write a script (e.g., Python) to scan all .md files and auto-populate referenced_in_docs to detect orphaned images
• [ ] Add docs/IMAGES.md with a searchable gallery (organized by category) linking to each tool's description and image, making visual discovery easier for new users

• Add a new section /docs/ChatGPT_limitations.md documenting known ChatGPT limitations (context window, knowledge cutoff, hallucinations) and workarounds; currently missing despite being fundamental for new users
• Create a /docs/ChatGPT_safety.md guide covering prompt injection risks, data privacy, and secure integration patterns for production use—increasingly important as more sensitive applications adopt ChatGPT
• Expand /docs/ChatGPT_tools.md with a structured comparison table (tool name, cost, integrations, use case) rather than just links; would make evaluation faster for new users

This is a documentation and resource collection repository with minimal security risk. It is primarily a curated list of ChatGPT tools, guides, and examples written in Markdown with supporting images and PDFs. No hardcoded credentials, database connections, or executable code was detected. The main risks are: (1) external link poisoning if upstream resources are compromised, (2) PDF/media file authenticity concerns, and (3) lack of dependency management (though minimal dependencies appear to exist). The repository follows responsible practices by including an MIT license and maintaining code on both GitHub and GitLab. Recommended actions focus on periodic auditing of external references and adding transparency around resource sourcing.

• Low · No Dependency Lock File — Root directory. The repository does not contain a package-lock.json, yarn.lock, or similar dependency lock file. This is a documentation/resource repository without apparent package dependencies, but if dependencies were added in the future, lack of lock files could lead to supply chain risks. Fix: If this project adds npm/yarn dependencies in the future, ensure lock files are committed to version control to prevent dependency version drift.
• Low · Public Resource Links Without Integrity Verification — docs/ directory (multiple markdown files). The repository contains links to external ChatGPT resources, tools, and projects in markdown files. If any of these external links become compromised or point to malicious content, users following these guides could be at risk. Fix: Periodically audit external links for validity and safety. Consider adding security warnings about verifying project authenticity before using third-party tools. Use HTTPS-only links.
• Low · PDF Files Without Source Verification — files/ directory. The repository includes PDF files in the files/ directory (e.g., AIGC video generation and Midjourney dictionary PDFs) without clear metadata about origin, verification, or checksums. Fix: Add metadata documenting the source, date, and SHA-256 checksums of all PDF files. Include warnings about verifying document authenticity before relying on their content.
• Low · Embedded Media Files Without Virus Scanning Metadata — docs/imgs/ directory. The repository contains numerous image files (GIFs, JPGs, PNGs) and at least one MP4 video file (chatgpt_engshell.mp4) without visible checksums or scan reports. Fix: Document any security scanning performed on media files. Consider adding checksums or digital signatures for verification.

LLM-derived; treat as a starting point, not a security audit.

• Open issues — current backlog
• Recent PRs — what's actively shipping
• Source on GitHub

Generated by RepoPilot. Verdict based on maintenance signals — see the live page for receipts. Re-run on a new commit to refresh.