# Setup posts

Blog posts about setup from looot.

## Posts (35)

- [Idempotency keys for paid APIs: never pay twice](https://looot.ai/blog/idempotency-keys-for-paid-apis) (2026-11-16): How an idempotency key stops a retry from charging you twice on a paid API, how to build one, and how to cap the cost of a call when a retry goes wrong.
- [LiteLLM for data APIs: fallback for your backend](https://looot.ai/blog/litellm-for-data-apis) (2026-11-16): LiteLLM for data APIs: how provider fallback works for lookups like email finding, with REST, Python and Node code and a cost cap per call.
- [NVD API with KEV and EPSS: how to triage a CVE](https://looot.ai/blog/nvd-api-kev-epss) (2026-11-16): Use the NVD API with CISA KEV and EPSS to triage a CVE: CVSS score, exploited status and exploit odds, plus NVD API key rate limits.
- [World Bank API in pandas, with Eurostat and ECB](https://looot.ai/blog/world-bank-api-pandas-eurostat-ecb) (2026-11-16): Call the World Bank API, Eurostat and the ECB from Python, turn three JSON formats into one pandas DataFrame, and see why their inflation numbers differ.
- [Claude Fable 5.2 MCP: setup and a real run with costs](https://looot.ai/blog/claude-fable-5-2-mcp) (2026-11-09): Claude Fable 5.2 MCP setup for Claude Code, claude.ai and the Claude API, plus a real 69-call looot run with the cost of every call.
- [AI Overview tracker: what 40 buyer queries show about who Google cites](https://looot.ai/blog/ai-overview-tracker) (2026-10-24): An AI Overview tracker in practice: we ran 40 buyer queries through Google AI Overviews. All 40 showed one. Here is who got cited, and how to track your own.
- [AI visibility tracker: check if ChatGPT names your brand](https://looot.ai/blog/ai-visibility-tracker) (2026-10-16): Build an AI visibility tracker with an agent: run buyer prompts through ChatGPT, Perplexity, Gemini and Google AI Overview, and log mentions and cited sources.
- [Web crawler vs web scraper: what each does, which you need](https://looot.ai/blog/web-crawler-vs-web-scraper) (2026-10-16): Web crawler vs web scraper: a crawler finds pages by following links, a scraper reads data from a page. What a web crawler is, robots.txt, and agent use.
- [Agent card: one-time virtual cards for AI agents, explained](https://looot.ai/blog/agent-card-one-time-virtual-cards-for-ai-agents) (2026-10-15): What an agent card is: a one-time or capped virtual card an AI agent pays with. Who issues them, limits, 3D Secure, and why single-use cards fail later.
- [Gemini 4 Argon API: key, first call and which tool to use](https://looot.ai/blog/gemini-4-argon-api) (2026-10-15): Gemini 4 Argon API how-to from Google's docs: get a paid key, make the first call with gemini-4-argon, the parameters that now fail, and grounding vs MCP.
- [Machine payments protocol: Stripe MPP explained for agents](https://looot.ai/blog/machine-payments-protocol-stripe-mpp-explained) (2026-10-15): The Machine Payments Protocol (Stripe MPP) explained: the 402 flow, shared payment tokens, Link's wallet for agents, and what a seller must build to accept it.
- [What is x402? HTTP 402 Payment Required for AI agents](https://looot.ai/blog/what-is-x402-http-402-payments-for-agents) (2026-10-15): What HTTP 402 Payment Required means, when servers send it today, and how x402 turns a 402 into a payment an AI agent makes inside the request.
- [Claude Code email finder without signing up to each provider](https://looot.ai/blog/claude-code-email-finder-without-signups) (2026-10-14): Use Claude Code as an email finder with one MCP connection and one prepaid balance, no account per provider. Setup, prices, and what a miss costs.
- [DeepSeek API web search: what works and what it ignores](https://looot.ai/blog/deepseek-api-web-search) (2026-10-14): DeepSeek API web search, from DeepSeek's docs: the Anthropic format, the new Search and Fetch APIs, what they cost, and what the Responses API ignores.
- [Gemini 4 Argon MCP: live web and company data](https://looot.ai/blog/gemini-4-argon-mcp) (2026-10-14): Gemini 4 Argon MCP setup: connect one remote MCP server to Google's new model for web search, page reading, company data and Google results.
- [How to get a TikTok transcript without an account](https://looot.ai/blog/how-to-get-a-tiktok-transcript-without-an-account) (2026-10-14): Get a TikTok transcript without a TikTok account. One looot command returns the captions as timestamped text. Real output, the price per call and the limits.
- [Instagram username search: three lookups for agents](https://looot.ai/blog/instagram-username-search) (2026-10-14): Instagram username search for agents: look up a public profile by username, turn a user ID into a username, or find accounts by keyword. Real outputs.
- [Muse Wonder MCP: give Meta's model live web data](https://looot.ai/blog/muse-wonder-mcp-live-data) (2026-10-14): Muse Wonder and MCP: Meta's API has no MCP field, so you add MCP servers in Muse Code, Claude Code or Codex. The config, the tool choice and the limits.
- [Pi coding agent MCP: add web search and lead data](https://looot.ai/blog/pi-coding-agent-mcp-data-tools) (2026-10-14): Pi coding agent MCP setup: add one remote MCP server to Pi 1.0 with pi mcp add, then search the web, read pages and find work emails.
- [YouTube transcript API: transcripts for your agent](https://looot.ai/blog/youtube-transcript-api-get-transcripts-for-agents) (2026-10-14): A YouTube transcript API for agents: one looot call returns timestamped captions and plain text. A real output, the price per call and the limits.
- [What is a SKILL.md? How agent skills work](https://looot.ai/blog/agent-skills-for-data-jobs) (2026-10-04): An agent skill is a SKILL.md file that tells your agent when to start a job, what to ask for and which calls to make. What one holds, with a real example.
- [Keyword research with Claude Code: a volume skill](https://looot.ai/blog/keyword-research-with-claude-code) (2026-10-04): Do keyword research from your terminal. A Claude Code skill checks volume, cost per click and difficulty for a keyword list and reads the price first.
- [Why does my AI agent make up emails? How to stop it](https://looot.ai/blog/why-your-agent-guesses-emails) (2026-10-04): An agent with no email tool will guess one from a pattern and sound sure. Why it happens, how to tell found from guessed, and the skill that stops it.
- [Claude Code for marketing: a project setup that works](https://looot.ai/blog/claude-code-for-marketing) (2026-10-03): How to set up Claude Code for marketing work: a project folder, a CLAUDE.md of brand rules, live data through one MCP connection, and three jobs to run first.
- [Claude Code MCP servers for data work: how to add and vet them](https://looot.ai/blog/claude-mcp-servers-for-data-work) (2026-10-03): Claude Code MCP servers for search, scraping, SEO and enrichment data: how to add them, what to check before you trust one, and how to keep the list short.
- [Claude RAG with live data from an API](https://looot.ai/blog/claude-rag-with-live-data) (2026-10-03): Claude RAG usually means a vector store. For facts that change, retrieving from a live API works better. How the two differ and how to wire Claude to live data.
- [How to build an MCP server, and when to skip it](https://looot.ai/blog/how-to-build-an-mcp-server) (2026-10-03): Build a small MCP server in about thirty lines of JavaScript, connect it to Claude Code, and learn when an existing server or gateway is the better choice.
- [Instagram scraper for agents: profiles and posts](https://looot.ai/blog/instagram-api-for-agents) (2026-10-03): An Instagram scraper for AI agents: public posts, reels and profile data through looot, with each call's price shown first and the limits spelled out.
- [LLM tool calling: how an agent calls a real API](https://looot.ai/blog/llm-tool-calling-with-live-apis) (2026-10-03): LLM tool calling lets a model ask your code to run a function and read the result. How the loop works, what breaks with real APIs, and how to keep it simple.
- [Reddit API for agents: posts and comments, one token](https://looot.ai/blog/reddit-api-for-agents) (2026-10-03): How to use the Reddit API from an AI agent through looot: posts, comments and subreddit search, what each call returns, and where the data stops.
- [TikTok scraper for agents: profiles and video stats](https://looot.ai/blog/tiktok-scraper-with-an-agent) (2026-10-03): Get public TikTok profile and video data into an AI agent through looot, with each call's price shown first. What a TikTok scraper returns and where it stops.
- [Twitter scraper for agents: search X posts with one token](https://looot.ai/blog/twitter-scraper-with-an-agent) (2026-10-03): How an AI agent searches X (Twitter) posts and people through looot, which providers answer, what a call costs, and what a Twitter scraper cannot give you.
- [What is an MCP server? A plain explanation](https://looot.ai/blog/what-is-an-mcp-server) (2026-10-03): An MCP server gives an AI agent tools it can call. What a tool call looks like, and how one connection can reach many providers.
- [YouTube scraper for agents: videos, views, channels](https://looot.ai/blog/youtube-scraper-with-an-agent) (2026-10-03): A YouTube scraper for AI agents: pull a channel's videos and view counts through looot, see what each call returns and costs, and where the data stops.
- [Set up looot with one line: skill.md and llms.txt](https://looot.ai/blog/set-up-looot-with-one-line) (2026-10-02): Give your agent one URL and it can set up looot itself. What skill.md and llms.txt are, what each one tells an agent, and a prompt you can paste today.

## Other categories

All posts: https://looot.ai/blog.md. RSS feed: https://looot.ai/blog/rss.xml

- [Comparisons](https://looot.ai/blog/category/comparisons.md), 31 posts
- [Costs](https://looot.ai/blog/category/costs.md), 24 posts
- [Setup](https://looot.ai/blog/category/setup.md), 35 posts
- [Product](https://looot.ai/blog/category/product.md), 17 posts
