ry0narushima

Bettersearch — AI skill for Claude Code

AI community

Deep research for any AI agent on DeepSeek's new Search API.

How to install Bettersearch

This entry records only its repository, not the path inside it, so there is no exact command to give. Open ry0narushima/bettersearch and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Bettersearch does

Deep research for any AI agent on DeepSeek's new Search API. Subagents, fact-checking, and a clean HTML report.

Alternatives in AI

Deepreasoning — A high-performance LLM inference API and Chat UI that integrates DeepSeek R1's CoT reasoning traces with Anthr 5.4k ★
Deepseek Pp — DeepSeek Web browser extension: AI agent workspace with MCP tools, memory, Skills, automation, web search, and 1.7k ★
Academic Commercialization Agent — Turn any research paper into a commercialization report — 6 AI agents, TRL/MRL scoring, patent landscape, mark 581 ★

README

BetterSearch

DeepSeek just opened up the search engine behind its app. Any agent can call it now for **$0.75 per 1,000 searches**.

Raw results are decent. But an agent sending one vague query gets one vague answer... and sometimes some of it in Chinese (we searched in English and 2 of the top 5 came back in Chinese).

BetterSearch is a skill that fixes the part around the API: better queries, parallel research, a fact-check pass, and a clean report at the end. It works in Claude Code, Codex, Cursor, Gemini CLI, OpenCode and Hermes.

Install

curl -fsSL https://raw.githubusercontent.com/ry0narushima/bettersearch/main/install.sh | bash

Then add your DeepSeek key, once:

python3 ~/.claude/skills/bettersearch/scripts/dsr.py setup

It opens the key page, tells you what to click, tests the key and saves it to `~/.config/deepseek-research/` (chmod 600). The key never goes through your agent's chat. `DEEPSEEK_API_KEY` works too.

Two modes

**Quick** is the default. Ask a normal question, get a sourced answer in chat.

you ──► 2-4 sharp queries ──► best 1-3 pages fetched ──► answer + [1] [2] citations
        (lang:, after:, site:)

**Advanced** kicks in when you say "deep research", or the question is big.

                        ┌──────────────┐
           question ──► │   planner    │  8-20 queries, up to 50 results each
                        └──────┬───────┘
                               ▼
                        ┌──────────────┐
                        │    gather    │  dedupe, rank by authority, fetch top 25
                        └──────┬───────┘
                ┌──────────────┼──────────────┐
                ▼              ▼              ▼
          ┌──────────┐   ┌──────────┐   ┌──────────┐
          │researcher│   │researcher│   │researcher│   one per sub-question,
          └────┬─────┘   └────┬─────┘   └────┬─────┘   in paral