它提供了大量真实的AI对话数据集（超过1.1万条消息）以及开源工具，支持增量归档、非拉丁语标题，并能处理Cloudflare防护的API。
An archive of public Claude, Grok, and Kimi conversations, exported from their share links as plain markdown, plus the three scripts that produce it. A handful of other conversations, added by hand from sources with no public share link, live alongside them.
1,334 conversations, 15,252 messages, ~38 MB.
| Directory | Source | Conversations | Messages | Longest |
|---|---|---|---|---|
claude_chats/ | claude.ai/share | 755 | 7,865 | 282 |
grok_chats/ | grok.com/share | 519 | 7,210 | 4,308 |
kimi_chats/ | kimi.com/share | 57 | 171 | 28 |
aistudio_chats/ | Google AI Studio (manual) | 3 | 6 | 4 |
scripts/claude_share_export.py exporter for claude.ai shares
scripts/grok_share_export.py exporter for grok.com shares
scripts/kimi_share_export.py exporter for kimi.com shares
scripts/aistudio_export.py converter for raw Google AI Studio page copies
link_lists/claudeaisharelinks.txt the claude.ai links this archive was built from
link_lists/grokaisharelinks.txt the grok.com links this archive was built from
link_lists/kimiaisharelinks.txt the kimi.com links this archive was built from
raw_google_dumps/ raw AI Studio page copies aistudio_export.py reads
claude_chats/ one markdown file per Claude conversation
grok_chats/ one markdown file per Grok conversation
kimi_chats/ one markdown file per Kimi conversation
aistudio_chats/ one markdown file per AI Studio conversation
claude_chats/.claude-share-export.json manifest for claude_chats/
grok_chats/.grok-share-export.json manifest for grok_chats/
kimi_chats/.kimi-share-export.json manifest for kimi_chats/
aistudio_chats/.aistudio-export.json manifest for aistudio_chats/
Every conversation is one markdown file, named from its title:
Adding-Kerberos-Authentication-to-Hadoop-Cluster-on-EC2.md
AI-inflection-point-and-the-future-of-automation.md
네이버-AI-모델서비스-기획-직무-기업조사-리포트.md
Titles keep their original script rather than being transliterated, so non-Latin
filenames are expected. A conversation with no usable title falls back to its
snapshot UUID (18754235-198d-446b-afc6-26191ea62d27.md).
All four converters write the same shape: a metadata header, then alternating turns separated by a
An archive of public Claude and Grok conversations, exported from their share links as plain markdown, plus the two scripts that produce it.
根据分类、Topic 和编程语言匹配的相似项目。
Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.
A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.
mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.