Hi HN!
Made this AICW Video open-source tool when had to quickly edit a number of raw video interviews.
It can:
- find which separate audio file belongs to which video and replace it (use case: audio recorded on a separate phone)
- auto crop a video (manual crop too)
- suggests clips to create
- adds dynamic captions
- exports to TikTok/Insta/others
- can blur a face, can overlay audio with tts
Uses local AI for audio to text, face detection. For scenes detection requires Claude or Codex but can use Ollama with Qwen. Mac only. Works standalone and as MCP via Claude/Codex.
HM looks great, but looks like they collect lot of personal data.
From their privacy policy:
The personal information we collect may include:
Name
User name
Email address
Gender
Birthday (month/year)
Ethnicity
Relationship Status
Number of Family Members/Children/Children under 13
Experience with Meditation
User generated content
We may collect other information that does not reveal your specific identity, such as:
IP Address or other unique device identifier
Information collected through cookies, pixel tags or other technologies
App usage data
Geo-location information
User generated content
Device generated data
SQLite3 is awesome but turned out if you create a new RoR application with table and child table with child records and then rename or remove a column then all child records are deleted. This issue is currently present in both 8.1.3 release and edge versions.
To grade AI visibility of your website one can use this command line:
npx @aichatwatch/aicw ai-visibility EXAMPLE.com
It checks how the server responds to AI crawler bots, analyzes the content (HTML) code, checks if the website is present in public datasets, and returns a final grade similar to this:
================================================================================
AI VISIBILITY REPORT for https://aichatwatch.com/ on Oct 26, 2025, 23:11 PM
================================================================================
SCORE: 86% (93.5/109 points)
GRADE: B
================================================================================
DETAILS:
[PASS] Server: check /robots.txt.................... 15/15
[WARN] Server: check /sitemap.xml................... 0/5
[PASS] Server: X-Robots-Tag HTTP Header Check....... 15/15
[PASS] Server Response Speed........................ 6.3/7
[WARN] Content: JSON-LD Structured Data Presence.... 0/5
[PASS] Content: Blocking Meta Tags Check............ 8/8
[PASS] Content: Server-Side Rendering (SSR) Check... 8/8
[PASS] Content: HTML code structure................. 4.2/5
[PASS] Content: Mobile Version Availability......... 7/7
[WARN] Dataset: /llms.txt available................. 0/1
[PASS] AI bots access: Data for Training............ 12/12
[PASS] AI bots access: Search Indexes............... 10/10
[PASS] AI bots access: User Interactions............ 8/8
[WARN] Dataset: Common Crawl Dataset................ 0/3
--------------------------------------------------------------------------------
ISSUES:
--------------------------------------------------------------------------------
Server: check /sitemap.xml
Impact: 5 points lost
Details: No /sitemap.xml found
Content: JSON-LD Structured Data Presence
Impact: 5 points lost
Details: No JSON-LD structured data found.
Dataset: /llms.txt available
Impact: 1 point lost
Details: No /llms.txt found (not critical)
Dataset: Common Crawl Dataset
Impact: 3 points lost
Details: Not found in 3 recent crawls: CC-MAIN-2025-43, CC-MAIN-2025-38, CC-MAIN-2025-33
There are some demo reports tracking who and how its mentioned on aichatwatch dot com, like: "Y Combinator", "Top AI Tools", "high blood pressure and hypertension", "Best Cities for Nomads"..
Hey HN! creator is here. AI Chat Watch is an open-source CLI tool that helps marketers to track and understand how AI chatbots talk about brands, products, industry.
It runs a set of questions across multiple AI models (currently ChatGPT, Claude, and Perplexity via OpenRouter) and generates reports showing:
- Which entities are mentioned, how often and in which order
- Trends of these mentions (comparing to previous dates)
- Specific links and domains cited in responses by AIs
Originally built this as a commercial product for marketers doing competitive research, but decided to open source it so anyone can run it locally. You'll need to provide your own OpenRouter API key (~$0.30 per run for 5 questions × 3 models).
Some example use cases:
- Competitive analysis (e.g., "best CRM software")
- Keywords monitoring across AI responses
- Exploring websites and specific pages cited by AI
- SEO/visibility in AI-generated answers
Installation is straightforward via terminal. There are demo reports in the repo to see what the output looks like.
Would love feedback and suggestions on making it more valuable
Thanks
Made interactive text based trainer for recognizing phishing emails, messages, documents, sms and others. Look at the message provided and answer if it is legit or not?
Runs on ChatGPT platform
why spam? this is an aggregator from few sites where owners are selling apps and websites. These apps are real, I've contacted owners of few app sold and they are legit.