Skip to content

Latest commit

 

History

History

README.md

🔎 OmniSearch

OmniSearch is a research assistant that reads the internet for you.

Ask it what people on Hacker News, Stack Overflow, GitHub, arXiv, Dev.to, Lobsters, and Bluesky are saying, asking, and publishing — and it brings back a clear answer with links.

You never need to touch the command line. This tool is for your AI agent (Claude Code, Codex, Cursor, ChatGPT), not for you. Just tell your agent what you want to learn, and it handles the rest.


How to use OmniSearch (no coding needed)

There are three steps — and you only do the first one yourself.

Step 1 — Install once (5 minutes). You add OmniSearch to your AI assistant, like installing an app once. Your agent runs this for you:

webcmd plugin install github:agentrhq/webcmd/omnisearch

Step 2 — Tell your agent what you want to know.

You don't type commands. You just talk. Say something like:

"Before I build this, find out what people actually think about it."

Your agent recognizes OmniSearch and runs it for you.

Step 3 — Read the answer.

You get a plain-language summary — what people say, where they say it, and links to the actual posts. No folders, no files, nothing to interpret.


What you say. What your agent does.

Here are real conversations, ready to copy-paste. Swap the words in {braces} for your topic.

You say What your agent finds
"{My idea} — do people actually want this?" What people say across Hacker News, GitHub, and developer communities
"Find what people complain about when using {a tool}." The real problems people report on Stack Overflow and GitHub
"Is {this idea} already done? Who's winning?" The competition and reception across platforms
"Is {this idea} even technically possible yet?" What research says is proven vs. still speculative
"Before I launch {product}, what could kill it?" The risks, past failures, and strongest arguments for it
"What's the reaction to {a product / a launch / the news}?" What people are saying on Hacker News, Lobsters, and Bluesky right now

Ready-to-paste prompts for your agent

Pick one, replace {the brackets}, and paste it to your agent:

  1. "Research {topic/competitor/product}. Use OmniSearch to check Hacker News, Stack Overflow, GitHub, arXiv, Dev.to, and Lobsters. Give me a two-paragraph summary and the 5 most-talked-about results with links."

  2. "Find what's broken or frustrating about {topic}. Look at Stack Overflow questions and GitHub issues, and list the 10 most common problems people mention, each with a source link."

  3. "Check if {my idea} is technically realistic yet. Scan recent research papers and tech discussions, and tell me what's proven to work vs. still speculative."

  4. "Validate my idea: {one-line pitch}. Gather what people are saying about it, rate how much real interest there is, and quote 3 real people with links."

  5. "Do a pre-mortem on {my startup/product}. Find how similar things failed before, the top 3 reasons this could fail, and the strongest argument it could succeed — each with a source."

  6. "What do developers complain about when using {a tool/language}? Summarize the biggest recurring pain points with links."

  7. "Check the reaction to {a new release / launch}. What are people on Hacker News, Lobsters, and Bluesky saying right now?"


What you can research

OmniSearch reads 7 public sources without requiring an account:

Platform Best at telling you
Hacker News 🟠 what the tech world thinks
Stack Overflow 📚 what's actually broken
GitHub 🐙 real problems people file
Dev.to 💜 what developers are writing
arXiv 📄 the latest research
Lobsters 🦞 smart developer discussion
Bluesky 🦋 fresh public posts

No login is needed. Each command uses a public API or feed.


Start now

Install OmniSearch once, then ask your agent:

"Before I build this, find out what people actually think about it."

That's the whole product.


For developers

Everything below is for the agent (and for people who want to dig in). A normal user never reads this.

The commands

# Aggregate everything about a topic
webcmd omnisearch research "<topic>" -f json

# 🏆 The community's verdict, distilled
webcmd omnisearch verdict "<topic>" -f json

# One source at a time
webcmd omnisearch hackermind "<q>" -f json
webcmd omnisearch stackoverflow "<q>" -f json
webcmd omnisearch github "<q>" -f json
webcmd omnisearch arxiv "<q>" -f json
webcmd omnisearch research <tag> --sources devto -f json
webcmd omnisearch lobsters -f json
webcmd omnisearch bluesky-posts <handle> -f json

Output

The research and source-search commands return platform, title, author, score, commentCount, createdAt, url, and text. Feed commands expose source-specific fields.

How it works

The aggregated research sources live in sources.js behind one signature (search(query, limit)).

Ships SKILL.md so agent harnesses auto-discover OmniSearch.