AI Search Hub
Sends one prompt to Doubao, Yuanbao, Qwen, Gemini or Grok via a scripted Chrome session
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jul 21, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · d5b32f8
I cloned the scripts and ran the skill's documented command for real: `python3 scripts/run_web_chat.py --site gemini --prompt "What is 2+2?" --headless --user-data-source <empty dir> --debug-profile-dir <tmp>` — it auto-started Chrome on 127.0.0.1:9222, opened Gemini, sent the prompt and wrote "2 + 2 = 4" to the --output file on the first try. For the baseline I first tried the naive path I would have taken without the body (`python3 scripts/gemini_playwright.py "What is 2+2?"`); it failed on a positional-arg error, then on "Standalone launch requires --profile-dir", and the third attempt with --profile-dir hung on Playwright click-retry against an overlay and produced no answer file — 3 attempts, 0 results vs 1 command, 1 answer. Verified flags: every argument documented in SKILL.md exists in the real argparse, ROUTING.md and scripts/run_web_chat.py return HTTP 200, but INSTALL.md referenced at SKILL.md line 31 returns 404 and `kimi` is a supported --site in code yet missing from the SKILL.md site list and routing table. No exfiltration or curl|sh; the wrapper does copy the user's browser user-data-dir (cookies/sessions) into a repo-local chrome_debug_profile_skill (gitignored) and exposes CDP port 9222 locally, and the README oversells with claims of "captcha, rate-limit, and anti-bot recovery" and link fetch/clean/structure that the CLI does not implement (login is manual wait; the only input is --prompt). Note: my first download into the shared scratchpad root was silently overwritten by another parallel process with a different skill's SKILL.md, so I re-fetched into a unique directory and re-verified by checksum.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 4/5
- Triggers reliably 5/5
- Output vs. baseline 8/10
- Docs & honesty 3/5
What AI Search Hub does
Drives the repository's Playwright scripts to send a prompt to a chosen AI chat site (Yuanbao, LongCat, Doubao, Qwen, Kimi, Gemini, Grok, MiniMax), attaching to or auto-starting a Chrome DevTools session on port 9222 and writing the reply to a text file. Includes a routing table that maps a question's data world to the platform most likely to answer it (WeChat content to Yuanbao, Douyin to Doubao, Meituan to LongCat, X to Grok). Triggers when the user asks to run a prompt through one of those sites, reuse a debug browser profile, or wait for a manual login before continuing.
How to install AI Search Hub
mkdir -p ~/.claude/skills
git clone --depth 1 https://github.com/minsight-ai-info/AI-Search-Hub.git ~/.claude/skills/ai-search-hub
# The whole repo is the payload — scripts/ is required, SKILL.md alone is not enough.
# Then: pip install playwright && playwright install chromium
# Run: python3 ~/.claude/skills/ai-search-hub/scripts/run_web_chat.py \
# --site gemini --prompt "What is 2+2?" \
# --repo-root ~/.claude/skills/ai-search-hub --output ~/answer.txt
# Most sites (doubao, yuanbao, longcat, qwen, grok) need a logged-in browser:
# omit --headless so the wrapper pauses for you to log in, then continues.
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger AI Search Hub
-
/ai-search-hubSends one prompt to Doubao, Yuanbao, Qwen, Gemini or Grok via a scripted Chrome session
It also activates on plain-language prompts like these:
-
Ask this prompt to Doubao and Qwen -
Attach to my Chrome debug session and query Gemini -
Send this question to Grok and MiniMax
Frequently asked questions
- Is the AI Search Hub skill free?
- Yes. The skill itself is free from minsight-ai-info/AI-Search-Hub. SkillProof publishes the install command and an independent test verdict at no cost.
- Does AI Search Hub work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 21, 2026. Verdict: Tested · Works. I cloned the scripts and ran the skill's documented command for real: `python3 scripts/run_web_chat.py --site gemini --prompt "What is 2+2?" --headless --user-data-source <empty dir> --debug-profile-dir <tmp>` — it auto-started Chrome on 127.0.0.1:9222, opened Gemini, sent the prompt and wrote "2 + 2 = 4" to the --output file on the first try. For the baseline I first tried the naive path I would have taken without the body (`python3 scripts/gemini_playwright.py "What is 2+2?"`); it failed on a positional-arg error, then on "Standalone launch requires --profile-dir", and the third attempt with --profile-dir hung on Playwright click-retry against an overlay and produced no answer file — 3 attempts, 0 results vs 1 command, 1 answer. Verified flags: every argument documented in SKILL.md exists in the real argparse, ROUTING.md and scripts/run_web_chat.py return HTTP 200, but INSTALL.md referenced at SKILL.md line 31 returns 404 and `kimi` is a supported --site in code yet missing from the SKILL.md site list and routing table. No exfiltration or curl|sh; the wrapper does copy the user's browser user-data-dir (cookies/sessions) into a repo-local chrome_debug_profile_skill (gitignored) and exposes CDP port 9222 locally, and the README oversells with claims of "captcha, rate-limit, and anti-bot recovery" and link fetch/clean/structure that the CLI does not implement (login is manual wait; the only input is --prompt). Note: my first download into the shared scratchpad root was silently overwritten by another parallel process with a different skill's SKILL.md, so I re-fetched into a unique directory and re-verified by checksum.
- What is the AI Search Hub SkillProof Score?
- 8.0/10 — installs cleanly 4/5, triggers reliably 5/5, output vs. baseline 8/10, docs & honesty 3/5.
- How do I install AI Search Hub?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use AI Search Hub with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.