07-11 14:38 · 开源情报,行业分析,趋势监控,决策支持,AI工具
Evidence-backed AI industry intelligence — trends, source updates, daily data refreshes, and weekly decision briefs.
English · 简体中文
Evidence-backed AI industry intelligence for people who need to decide, not just keep up. Agent Pulse turns official releases, research papers, capital moves, policy changes, and public propagation signals into traceable Events, evolving industry judgments, and explicit next signals.
Open Agent Pulse · ⭐ Star this repository · Read in Chinese
AI has no shortage of news, feeds, or hot takes. What is scarce is a reliable way for investors, executives, founders, business leaders, and technical leaders to answer three questions:
- What materially changed? Not every announcement deserves a new view.
- Why does it matter? A fact becomes useful only when connected to an industry shift and the people it affects.
- What would change the judgment? Every durable view needs a next signal, counter-signal, or invalidation condition.
Agent Pulse is built around that loop: observe broadly, publish selectively, preserve the evidence trail, and update a judgment only when the evidence changes.
If that is the kind of public intelligence infrastructure you want to exist, star Agent Pulse. It helps more decision-makers and builders discover the project and is the clearest signal that the open version is worth maintaining.
Every public event is designed to separate six layers:
- Fact — what happened and when;
- Evidence — the original release, paper, filing, or policy document;
- Context — what came before and what actually changed;
- Impact — who is affected across technology, business, capital, and policy;
- Judgment — the current interpretation, with uncertainty made explicit;
- Next signal — what would strengthen, weaken, or invalidate the judgment.
official releases + papers + filings + expert and propagation signals
│
▼
collect → normalize → deduplicate → cluster → evidence binding
│
▼
deterministic gate → strategic narrative → next signal → public event
Aggregators can suggest candidates or propagation heat, but cannot become the sole factual evidence for a material event. External text is never treated as trusted HTML, and private operational data never enters the public static export.
The repository is not a mockup. It contains the source catalog, collectors, evidence model, automatic quality gates, static renderer, source-health automation, and GitHub Pages delivery path used by the live product. It also keeps four states deliberately separate: catalogued, observed, active, and published.
GitHub Actions refreshes public data and redeploys the static site once per day. The source audit, monitor, and quality guard remain weekly; the weekly-brief
Issue is created or updated only on Sunday (or by an explicit manual run) and only when at least one public Event clears the weekly gate, so daily freshness does not create empty Issues.
Repository evidence checked on 2026-07-14. Source-health numbers below come from the full audit completed at 2026-07-14 05:36 UTC; content counts come from the versioned repository snapshot.
See the machine-generated source health report, system capability evaluation, data-source policy, and capability map. CI restores the complete repository snapshot, compares a fresh evaluation with the previous commit, fails on score or evidence regression, and publishes the highest-priority evidence gaps as its improvement queue.
The public source-update stream exposes only allowlisted titles, attribution, dates, categories, tags, and canonical links. It remains an observation layer and cannot bypass the evidence and readiness gates required for Event publication. The limitations are equally important: production qualification still needs a real observation window; many historical events need more independent evidence; claim-level evidence, multilingual semantic clustering, real MySQL integration coverage, and user outcome feedback are still incomplete. Planned or experimental capabilities are never presented as shipped.
sources and discovery signals
│
SourceAdapter boundary
│
safe fetch → normalize → quality gate → deduplicate
│
isolated observation / event clustering
│
evidence binding → deterministic readiness gate
│
┌────┴────┐
▼ ▼
Control Room privacy-safe static site → GitHub Pages
SQLite is the zero-configuration default. A MySQL dialect path exists, but the project does not claim MySQL compatibility without real integration coverage. Public Pages contain only allowlisted DTOs; databases, credentials, raw payloads, proxy settings, and private notes are excluded.
Read the architecture, roadmap, or changelog for implementation details and release history.
Requires Node.js 22 or later.
git clone https://github.com/barretlee/agent-pulse.git
cd agent-pulse
npm install
cp .env.example .env
npm run dev
Local startup automatically migrates the database, refreshes catalog seed metadata, and merges the latest versioned repository snapshot. A fresh clone and an existing local SQLite database therefore start from the same complete repository dataset without deleting newer local evidence. npm run db:
▸ 展开全文
07-02 17:33 · AI代理,量化交易,开源,MCP,金融科技
AI-assisted trading desk for short-term technical analysis on stocks & ETFs via Robinhood MCP. Deterministic Python engines score each asset on a three-pillar framework (Trend · Momentum · Macro-Senti
AI-assisted trading desk for short-term technical analysis on stocks & ETFs via Robinhood MCP. Deterministic Python engines score each asset on a three-pillar framework (Trend · Momentum · Macro-Senti
▸ 展开全文
06-28 20:36 · GitHub,开源,技术,Go,技术
PhonePe Transaction Trends & User Analytics Dashboard 2026 – Power BI Guide
PhonePe Transaction Trends & User Analytics Dashboard 2026 – Power BI Guide
▸ 展开全文
06-28 20:28 · GitHub,开源,数据可视化,Netflix
Top Netflix Movie & Show Trends Dashboard 2026 – Power BI Data Analytics
Top Netflix Movie & Show Trends Dashboard 2026 – Power BI Data Analytics
▸ 展开全文
06-28 20:24 · 开源,金融科技,量化交易,实时信号,外汇
MT5 Live Trend Direction Signal Hub 2026 – Real-Time Forex & CFD Alerts
MT5 Live Trend Direction Signal Hub 2026 – Real-Time Forex & CFD Alerts
▸ 展开全文
06-28 20:22 · 开源,算法交易,金融科技,MT5,GitHub Trending
2026 MT5 Gateway: Institutional Trend Classifier & Multi-Asset Regime Dashboard
2026 MT5 Gateway: Institutional Trend Classifier & Multi-Asset Regime Dashboard
▸ 展开全文
06-28 19:21 · 开源,数据分析,电商仪表板,Go
Top 5 Amazon Sales Dashboard Trends & Insights for 2026
Top 5 Amazon Sales Dashboard Trends & Insights for 2026
▸ 展开全文
06-22 20:30 · 加密货币,开源,API集成,macOS工具,数据可视化
CryptoBar brings real-time cryptocurrency prices to your macOS menu bar using the CoinGecko API. With built-in 24-hour charts, it lets you instantly track market trends without opening a separate app.
Real-time cryptocurrency prices in your macOS menu bar
CryptoBar is a lightweight macOS menu bar application that displays real-time cryptocurrency prices using the CoinGecko API. It also shows 24-hour price charts for quick market trend overview directly in the menu bar.
- Real-time cryptocurrency prices in macOS menu bar
- 24-hour price charts for quick trend visualization
- Lightweight and fast background operation
- Powered by CoinGecko API
- Minimal UI with focus on data clarity
- Download the latest release: https://github.com/erdwin90/crypto-bar-mac/releases/latest
- Open the .dmg file and install the app
- Launch CryptoBar — it will appear in your menu bar
- Click the menu bar item to see detailed prices
CryptoBar uses the public CoinGecko API and does not require an API key.
You can customize:
- tracked coins
- update interval
- UI layout
- Objective-C
- AppKit
- macOS Menu Bar (NSStatusItem)
- CoinGecko API
MIT License. See LICENSE for details.
If you encounter any issues or have feature requests, please open an issue on GitHub.
▸ 展开全文
06-18 16:06 · 开源,社交媒体监测,趋势分析
Twitter Trend Radar
English | 简体中文
Scan Twitter/X in real time for tweets that carry links and have traction, then reverse-look up each domain's registration age and recent traffic — so you can spot emerging products, tools, and sites before Google Trends catches up.
The idea: Google Trends is a lagging indicator. Something blows up on social first, that drives people to search, and only then does the Trends curve start to rise. By the time you see it there, the early window is gone. The source lives on X — pull the launch-moment tweets ("launch signal phrases" + links + engagement + a time window), reverse-check domain age and traffic, and you're a step ahead.
- Keyword pool with round-robin — ships with a set of launch-signal phrases
(
just launched
/introducing
/built this
/made a site
…). "Scan all" runs each once; auto-patrol rotates through them, so you're continuously watching X for "someone just shipped something". - Precise filtering — X advanced-search
filter:links
+ a minimum-likes floor + a time window (since:
). Only recent, link-bearing, engaged tweets. - Greylist — automatically skips major sites (google / youtube / github / openai …) that can never be the "new site" you're hunting.
- Domain age — whois lookup tags each domain new / recent / old. TLDs that don't expose
a creation date (many
.ai/.io/.app
) are marked age unknown and are not discarded — traffic decides instead. - Recent traffic — a bar chart of the last three months, so a ramp (e.g. 0 → 0 → 2M) is obvious at a glance.
- Platform subdomains —
xxx.vercel.app
/xxx.lovable.app
etc.: the root domain's age is meaningless, so it falls back to traffic automatically. - Instant re-filter — all results stay in the frontend; changing the "max domain age" shows/hides cards instantly without re-querying.
- Local proxy — your keys live only on your machine, the web page never sees them, CORS is bypassed, and screen recordings won't leak anything.
Browser (index.html) ──> Local proxy (server.py) ──> Three external APIs
radar console UI holds keys / bypasses CORS ├─ AISA Twitter search
live render / filter whois + traffic + cache ├─ query.domains domain reg. date
└─ aitdk domain traffic
server.py
— a zero-dependency local HTTP server that is both a static server (serves the page) and an API proxy.index.html
— single-file frontend, no build step.
It runs with just the AISA key — you'll simply have no traffic charts, and domains show "age unknown".
Option A: .env file (recommended)
cp .env.example .env
# edit .env and fill in your keys
Option B: environment variables
export AISA_API_KEY="your_key"
export AITDK_API_KEY="your_key" # optional
export QUERY_DOMAINS_KEY="your_key" # optional
python3 server.py
Open the URL printed in the terminal (default http://127.0.0.1:8787) and start scanning.
Requires Python 3 (built-in is fine). No pip install needed.
- Keyword pool — type a term and press Enter, click a launch-signal chip below, or "Add all".
- Filters — minimum likes, links-only, tweet time window (last 1 month … 2 years), max domain age.
- Scan all — queries every keyword in the pool once.
- Auto-patrol — rotates one keyword every N seconds, continuously.
- Matching tweets stream in; age tags and traffic charts fill in progressively. A green border = new site (≤30 days) or a platform subdomain whose traffic is rising.
- Change "max domain age" anytime — results adjust instantly, no re-query.
The greylist and deployment-platform suffix list live at the top of server.py
(GREYLIST
/ PLATFORM_SUFFIXES
) — edit them as you like.
- All keys are read from environment variables /
.env
. No keys are stored in the code. .env
is ignored by.gitignore
, so it won't be committed by accident.- Keys live only in the local proxy process; the web page and browser never see them.
- If you fork and modify the code,
git grep
once before committing to make sure no key slipped in.
- This project is for retrieval and learning over public information only. Respect each API provider's terms of service and X's usage policy.
- Traffic and registration-date data come from third-party APIs and are for reference only.
MIT
▸ 展开全文
06-08 08:50 · 开源,视频生成,智能体,内容自动化,GitHub趋势
Trend-to-Video Agent Framework for publish-ready content packages
Trend-to-Video Agent Framework
Turn trends into publish-ready short video packages with trend intelligence, creator fit, viral prediction, and quality control.
Quick Start · What It Builds · Architecture · Output Example · Why Different
Trend2Video Pro is a Trend Intelligence + Content Execution System.
It is not a dashboard.
It is not a simple AI video generator.
It is an execution-first agent framework that decides what to make, generates the assets, checks quality, and exports a ready-to-review content package.
Input trend -> system decides -> system generates -> publish-ready package
The final artifact is a local folder, not only a video file.
flowchart LR
A["Trend"] --> B["Scoring"]
B --> C["Creator Fit"]
C --> D["Script"]
D --> E["Fact Check"]
E --> F["Storyboard"]
F --> G["Video"]
G --> H["Quality Review"]
H --> I["Publish Package"]
Agent layer:
src/agents/
trend_scout.py
trend_analyzer.py
creator_strategy.py
script_writer.py
fact_checker.py
storyboard.py
video_producer.py
quality_reviewer.py
orchestrator.py
Main entry point:
from src.agents.orchestrator import run_trend_to_video
result = run_trend_to_video(
topic={"title": "AI Agent Trend", "url": ""},
platform="Bilibili",
style="Tech News",
duration=60,
)
outputs/publish_packages/20260613_222144_ai-agent-trend/
video.mp4
title.txt
description.txt
hashtags.txt
thumbnail.png
subtitles.srt
quality_report.md
metadata.json
This package is the product output: a creator can review it, edit copy if needed, and publish manually.
Trend2Video Pro focuses on the full execution loop, not one isolated generation step.
The Streamlit UI has only three sections:
1. Input: trend/url + platform + style + duration
2. Execute: one-click agent pipeline
3. Output: video preview + score + package downloads
No analytics dashboard. No charts. No raw logs. The UI is built for creator execution.
git clone https://github.com/2417467487-hub/Trend2Video-Pro.git
cd Trend2Video-Pro
python -m venv .venv
Windows:
.venv\Scripts\activate
macOS/Linux:
source .venv/bin/activate
Install dependencies:
pip install -r requirements.txt
playwright install chromium
copy .env.example .env
macOS/Linux:
cp .env.example .env
Run the product UI:
streamlit run app.py
python main.py generate --title "AI Agent Trend" --platform "Bilibili" --style "Tech News" --duration 60
python main.py update-topics
python main.py list-topics
python main.py generate-from-topic --topic-id 1 --platform "Xiaohongshu" --style "Tech News" --duration 60
uvicorn api:app --reload
Endpoints:
POST /api/generate
POST /api/score
GET /api/topics
POST /api/update-topics
POST /api/generate-from-topic?topic_id=1
GET /health
Creator memory lives in src/creator/
and stores:
- niche
- tone
- target platforms
- audience
- past content memory
- creator-topic fit signals
It influences scoring and content angle recommendation before generation.
The prediction engine lives in src/prediction/
and outputs:
viral_probability
predicted_views
confidence_level
explanation
Inputs include trend score, competition, urgency, creator fit, and hook strength. The current baseline is transparent and rule-based, with room for an sklearn model once enough history exists.
Quality checks are implemented in code:
- 3-second hook signal
- 3 key points
- CTA presence
- hook score
- clarity score
- density score
- factual risk score
- overall score
Scripts below quality threshold are rewritten once before the package is finalized.
python evaluation/run_benchmark.py
Mock-mode benchmark metrics:
- hook score
- viral accuracy proxy
- script quality
- publish readiness
pytest
The test suite runs without API keys.
- Citation-aware fact checking
- More creator profile presets
- More visual templates and subtitle styles
- Trained viral prediction once enough package history exists
- Optional publishing integrations after local export is stable
MIT. See LICENSE.
▸ 展开全文