Multi-AI Feed scans five content platforms for AI trends, ranks viral topics, performs cross-source research, and delivers structured daily reports.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install multi-ai-feed
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install multi-ai-feed using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Multi-AI Feed?
Multi-AI Feed is an Openclaw Skills research and trend-monitoring workflow for Kuaishou, WeChat Official Accounts, Bilibili, Channels, and Xiaohongshu. It can query one platform, several platforms, or all five at once, filter content using platform-specific engagement metrics, cluster related topics, and compare AI trends across platforms.
The skill produces a visual HTML daily report with platform tabs, cross-platform topic comparisons, engagement statistics, cover images where available, categorized content, and AI intelligence reports for the top topics. It also supports normalized JSON output, custom keyword monitoring, source credibility labels, multi-engine verification, and daily subscriptions for repeatable AI market and content research.
Multi-AI Feed Use Cases
- Discover viral AI tutorials, tools, models, and AI-generated content across five major Chinese platforms.
- Create a daily multi-platform AI news, trend, or intelligence report from one CLI command.
- Compare how the same AI topic performs on Bilibili, Xiaohongshu, Kuaishou, WeChat, and Channels.
- Monitor focused themes such as RAG, vector databases, fine-tuning, LoRA, ComfyUI, Stable Diffusion, ChatGPT, Claude, or AI agents.
- Support competitor intelligence, product research, market monitoring, reputation tracking, and creator discovery.
- Generate machine-readable JSON for downstream agents, dashboards, data pipelines, or secondary analysis.
- Install a daily subscription that saves the selected platforms and keywords for scheduled reporting.
How Multi-AI Feed Works
- Validate the requested date against each platform's data availability schedule before making any API request. Kuaishou and Bilibili update at 15:00 daily; WeChat Official Accounts, Channels, and Xiaohongshu update at 16:00.
- If the requested date has no data, inform the user of the latest available date and wait for explicit confirmation before using
--latest. If the user declines, no data interface is called. - Resolve the API key, platform aliases, date range, keywords, result count, and output options.
- Query the selected platforms through their supported batch or per-keyword interfaces, applying pagination and stopping when the requested
--countis reached. - Normalize results across platforms and remove duplicates using
photoId. - Rank content using each platform's primary metric: interaction totals for Kuaishou, WeChat, Channels, and Xiaohongshu, and
likeCountfor Bilibili. - Extract tags and cluster related content into adaptive topic groups, splitting large groups by secondary tags and frequent title terms until useful categories are produced.
- Sort categories and content by volume and platform engagement, limiting each category to
--top-nitems, with a default of five. - Select the top three topics from each platform and automatically assign an intelligence mode: competitor research, public-opinion investigation, person background research, or information cross-validation.
- Search multiple engines in three rounds—broad scanning, deep investigation, and cross-source verification—and attach credibility labels to findings.
- Render the normalized results, cross-platform comparison, category summaries, investigation reports, and platform-specific sections into an HTML report, while optionally returning normalized JSON.
- Save the report under the configured output directory, open it in a browser unless
--no-openis specified, and optionally install or manage a daily subscription.
Multi-AI Feed Setup
1. Configure the API key
Obtain a RedFoxHub API key in the format ak_xxxxxxxx, then configure it using one of the supported priority levels: the --api-key argument, the REDFOX_API_KEY environment variable, a shell startup file, or ~/.qoder/apis/redfox.json.
macOS/Linux:
export REDFOX_API_KEY="ak_xxxxxxxx"
source ~/.zshrc # or source ~/.bashrc
python3 -c 'import os; print(bool(os.environ.get("REDFOX_API_KEY")))'
Windows PowerShell:
[Environment]::SetEnvironmentVariable("REDFOX_API_KEY", "ak_xxxxxxxx", "User")
2. Run the skill
Use the bundled entry point at $SKILL_PATH/scripts/multi_ai_feed.py. The default platform selection is all.
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" --platform all --latest
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" --platform bili,xhs --latest
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" --platform kuaishou --date 2026-08-30
3. Customize output and research
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" \
--platform all \
--keywords "RAG,ComfyUI,AI绘画" \
--latest \
--json
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" \
--platform all \
--latest \
--no-intel \
--no-open
Use --platform-keywords for per-platform overrides. Platform-specific keywords take precedence over global --keywords, which take precedence over built-in defaults.
4. Enable daily delivery
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" \
--platform bili,xhs \
--keywords "AI绘画" \
--subscribe
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" --show-subscription
python3 "$SKILL_PATH/scripts/multi_ai_feed.py" --unsubscribe
The default report directory is ~/Downloads/QoderReports. Use --output-dir to change it. Exit code 0 indicates success, 1 indicates a missing key or subscription failure, and 2 indicates invalid arguments.
Multi-AI Feed Data Schema & Taxonomy
Platform taxonomy
| Platform | Accepted identifiers | Primary ranking metric | Update time | Time filtering |
|---|---|---|---|---|
| Kuaishou | kuaishou, ks |
Likes + shares + comments | 15:00 daily | Supported |
| WeChat Official Accounts | gzh, wechat |
Likes + shares + comments | 16:00 daily | Not supported by the interface |
| Bilibili | bili, bilibili |
likeCount |
15:00 daily | Supported |
| Channels | sph, channels |
Likes + shares + comments | 16:00 daily | Supported |
| Xiaohongshu | xhs, xiaohongshu |
Likes + shares + comments | 16:00 daily | Supported |
Identifiers accept Chinese names, English names, and aliases without case sensitivity. --platform all selects every platform.
Report artifacts
- HTML report:
~/Downloads/QoderReports/AI多平台日报_{platform}_{date}.html. - Platform identifier:
allfor all platforms, a+-joined alias such asbili+xhsfor multiple platforms, or a single alias for one platform. - JSON mode:
--jsonwrites normalized structured results to stdout and sends progress logs to stderr. - Subscription state: stores the current
--platformand keyword configuration for scheduled daily execution.
Normalized report sections
- Cross-platform overview with platform, category count, item count, leading metric, and total interaction.
- Cross-platform topic comparison showing topic coverage and platform distribution.
- Per-platform category summaries with counts, percentages, and notable content.
- Emerging-topic signals based on relatively low volume and high engagement.
- Core creator or account statistics where available.
- Investigation reports containing dimensions, findings, sources, credibility, and key conclusions.
- Cross-platform comparison recommendations.
Credibility taxonomy
| Label | Meaning | Evaluation rule |
|---|---|---|
| ✅ Confirmed | Reliable information | Two or more independent sources agree |
| ⚠️ Pending verification | Disputed information | Sources conflict or evidence is incomplete |
| ❌ Refuted | Unreliable information | An authoritative source contradicts it |
| 🔍 Single source | Requires further validation | Only one source is available |
Source grades range from A-level official, government, or authoritative media sources to D-level anonymous or unverified sources. The report template preserves all required sections even when a module has no data, using an explicit no-data notice instead.
Multi-AI Feed Advanced Features
- Multi-platform orchestration supports one, several, or all five platforms through a unified CLI.
- Platform aliases are normalized centrally, and new platforms can be added by extending the
PLATFORMStable infeed_core.py. - Keyword precedence supports platform-specific overrides, global keyword sets, and built-in platform defaults.
- Query efficiency combines batch requests, per-keyword requests, pagination limits, deduplication by
photoId, and early stopping at--count. - Adaptive topic clustering extracts tags, splits oversized groups by secondary tags and title frequency, merges small groups, and limits category output with
--top-n. - Automatic investigation-mode selection covers competitor intelligence, public-opinion events, person background research, and information cross-validation.
- Multi-engine research can use Baidu, Google, WeChat, Toutiao, DuckDuckGo, Brave, Startpage, Google Scholar, WolframAlpha, and other configured engines according to the investigation goal.
- Credibility-aware reports distinguish confirmed, disputed, refuted, and single-source claims instead of presenting all search results as equally reliable.
- The HTML template includes platform filtering tabs, topic colors, cross-platform comparisons, statistics, investigation sections, and browser launching.
--jsonprovides agent-friendly structured output while keeping operational logs on stderr.--no-inteldisables the automatic intelligence investigation phase, and--no-openprevents automatic browser launching.- Subscription commands provide install, inspection, and removal of daily scheduled tasks; the standard daily delivery time is 17:00 local time.
- Rate-limit handling retries response code
3108once after five seconds, while invalid key codes3106and3107stop execution with a configuration prompt. - The workflow includes layered date-availability safeguards in both the agent lifecycle and
main(), preventing unintended API calls and avoiding credit consumption when data is unavailable.
SKILL.md
Loading
Related Openclaw Skills
WeChat Channels Ad Audit
A structured WeChat Channels creator audit skill that turns account, content, and engagement data into investment-ready advertising reports.
Skill Creation Architect
A quality-gated framework for designing, optimizing, auditing, and preparing AI agent skills for delivery.
CONNSKILL Growth Services
CONNSKILL Growth Services gives agents pay-per-call access to SEO intelligence, SERP data, verification services, EU-hosted AI and x402 seller trust checks using USDC on Base.
html-shot
html-shot uses headless Chromium and Playwright to convert HTML, URLs and SVG designs into production-ready images and complete icon sets.
Alibaba Cloud ROS Agent
Securely connect to Alibaba Cloud ROS Agent through StartChat for guided remote infrastructure conversations, architecture planning, permissions, and deployment workflows.
Orthopedic Deal Matching Platform
A local CLI platform that securely matches orthopedic buyers and suppliers using five-dimensional scoring and mutual consent before contact exchange.
Featured*
