ahkohd

ahkohd / yagami

Public

A local-first web search agent

11
1
100% credibility
Found Mar 28, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Yagami is a local command-line web search agent that uses AI to discover, fetch, and summarize relevant web content.

How It Works

1
🔍 Discover Yagami

You hear about Yagami, a helpful tool that searches the web for you using smart thinking right on your computer.

2
📥 Get Yagami

Download and set up Yagami easily so it's ready to use on your machine.

3
🚀 Start the Helper

Turn on Yagami's quiet background service that waits for your questions.

4
🧠 Connect Smart Thinking

Link Yagami to your local AI helper so it can understand and find answers.

5
💭 Ask Your Question

Type any question like 'latest news on AI' and Yagami springs into action searching the web.

6
📄 See Clean Answers

Yagami shows you neat summaries from the best pages, with links to read more.

7
Need Deeper Info?
Quick Search

Get fast everyday answers.

📚
Deep Dive

Let it research thoroughly and check back for a full report.

🎉 Always Informed

Now you have a personal web researcher that gives you trustworthy info anytime, fast and private.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 11 to 11 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is yagami?

Yagami is a local-first web search agent built in TypeScript that runs a daemon on your machine for agentic searches using local LLMs like those from Ollama or vLLM. It handles queries via CLI commands like `yagami search "latest browser automation"`, `fetch` for page extraction, and `deep start` for async multi-hop research, all without cloud dependencies. Developers get privacy-focused web access with HTTP API endpoints and MCP support for plugging into other agents.

Why is it gaining traction?

In a world of cloud-locked AI tools, yagami stands out as a true local-first web app that pairs your own LLMs with headless browsing via Lightpanda CDP and search engines like DuckDuckGo. The agentic flow—planning searches, browsing, extracting markdown—delivers collated results or deep reports with citations, plus streaming progress and JSON output for scripting. Its Pi package integration hooks devs building local-first web agents without vendor lock-in.

Who should use this?

AI agent builders chaining local LLMs for research tasks, local-first webdev teams needing offline web intel, or solo devs prototyping with commands like `code "React hooks best practices"` or `company "Stripe"`. Ideal for privacy-conscious workflows avoiding API quotas.

Verdict

Promising starter for local-first web search agents (11 stars, 1.0% credibility), with solid CLI/docs but early maturity—expect tweaks for edge cases. Try if you're in local-first web development; skip for production without more testing.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.