Custom, realtime RSS feeds for Hacker News
This service provides RSS, Atom, and JSON Feed endpoints for Hacker News content. All feeds are generated on the fly from the Algolia HN Search API — no cache, no database.
Append a suffix to any endpoint URL to select the feed format:
| Suffix | Format | MIME type |
|---|---|---|
(none or .rss) | RSS 2.0 | application/rss+xml |
.atom | Atom 1.0 | application/atom+xml |
.jsonfeed | JSON Feed 1.0 | application/json |
/newest
Newest stories and polls
/newest/ai NEW
Newest stories matching AI-related keywords (see below)
/newest/noai NEW
Newest stories with AI-related stories filtered out
/frontpage
Current front page stories
/ask
Ask HN threads
/show
Show HN threads
/polls
Polls
/jobs
Job postings
/newcomments
Newest comments site-wide
/bestcomments
Best comments (scraped from HN best comments page)
/item id=
Comments on a specific story — requires id=<item_id>
/user id=
All submissions and comments by a user — requires id=<username>
/threads id=
Comments only by a user — requires id=<username>
/submitted id=
Submissions only by a user — requires id=<username>
/replies id=
Replies to a user or item — requires id=<username>
/favorites id=
User's public favorites — requires id=<username>
/classic
HN Classic page
/best
HN Best page
/active
HN Active page
/invited
HN Invited page
/pool
HN Pool page
/launches
HN Launches page
/whoishiring
All posts from the latest "Ask HN: Who is hiring?" thread
/whoishiring/jobs
Job offers only (seeking employees)
/whoishiring/hired
Candidates only (seeking employers)
/whoishiring/freelance
Freelance posts only
| Parameter | Description | Example |
|---|---|---|
q | Full-text search query | q=golang |
id | Username or item ID (required by some endpoints) | id=pg |
points | Minimum points filter | points=100 |
comments | Minimum comment count filter | comments=10 |
count | Number of items to return (max 100, default 20) | count=50 |
description | Set to 0 to omit item descriptions from feed entries | description=0 |
link | Set to url to link directly to the article instead of HN comments | link=url |
author | Filter by submitter username | author=dang |
search_attrs | Algolia search attributes (comma-separated: title, story_text, comment_text, url, author) | search_attrs=title |
/newest/ai and /newest/noai filter the newest stories feed based on AI content detection.
Matching is case-insensitive and treats hyphens, slashes, and dots as word boundaries
(so LLM-based matches llm).
The following terms trigger a match:
| Category | Terms |
|---|---|
| Concepts | artificial intelligence, machine learning, deep learning, neural network, large language model, generative ai, natural language processing, foundation model, diffusion model, retrieval augmented, prompt engineering, fine tuning, finetuning, vector database, embedding model, agentic, ai agent, ai model, ai system, ai tool, vibe coding, attention mechanism |
| Acronyms (whole-word) | AI, LLM, LLMs, AGI, RLHF, NLP, GenAI |
| Companies | OpenAI, Anthropic, DeepMind, Mistral, Cohere, Hugging Face, Stability AI, Inflection AI, xAI |
| Products / Models | ChatGPT, GPT-4, GPT-3, Claude, Gemini, Llama, Copilot, Midjourney, DALL-E, Stable Diffusion, Whisper, Grok, Mixtral, Perplexity, Devin, Cursor, Phi-3, Phi-4, o1, o3 |
/newest/ai fetches up to 3× the requested count from Algolia and
then filters client-side, so fewer items may be returned than count requests when AI
content is sparse. The same applies to /newest/noai when AI content is dense.