kagi-fastgpt

What This Skill Does

The kagi-fastgpt skill provides OpenClaw with the ability to retrieve synthesized, AI-generated answers by querying the Kagi search index. Unlike traditional search engines that return a list of links, this skill leverages the Kagi FastGPT API to process multiple web sources and condense them into a single, cohesive, and fact-checked response. It includes citations and references, making it an excellent tool for users who require authoritative information quickly without the overhead of manually reviewing individual search results. The skill is powered by an efficient Go binary, ensuring minimal latency and high performance during execution.

Installation

Installation is streamlined through the OpenClaw hub. Run the following command in your terminal to integrate the skill:

clawhub install openclaw/skills/skills/joelazar/kagi-fastgpt

Before running the skill, ensure your Kagi API token is configured. Add the following to your shell profile (~/.zprofile or ~/.bashrc): export KAGI_API_KEY="your-api-key-here"

Use Cases

Use this skill when you need direct, reliable answers to factual questions rather than raw research data. It is particularly effective for:

• Resolving quick technical queries (e.g., "How do I install a Go package?")
• Investigating recent events or current news.
• Fact-checking information where authoritative sources are required.
• Summarizing complex topics that span across multiple web documents.

Example Prompts

1. "What are the primary differences between PostgreSQL and MySQL for high-concurrency environments?"
2. "Summarize the latest developments in the Rust programming language as of this month."
3. "Explain how to clear the DNS cache on a Linux machine using systemd-resolved."

Tips & Limitations

• Pricing: Keep in mind this is a paid service (1.5¢ per query). Monitor your API usage to avoid unexpected billing.
• Precision: While FastGPT is excellent for synthesis, verify critical data against the provided citations if you are performing professional research.
• Options: Use the --no-cache flag if you suspect the information might be outdated and require the most recent index update, or --json if you are piping the output into another automated pipeline or script.