SerpKite
Google search API built for AI agents
The story
Overview
SerpKite is a Google SERP API purpose-built for AI agents and language models. It fetches live Google search results and returns them in LLM-optimized formats (JSON or Markdown), including all major SERP features like organic results, knowledge graphs, answer boxes, and People Also Ask in a single API call. The service is designed to reduce token usage by offering compact output formats and supports multiple search verticals including news, images, videos, maps, shopping, scholar, and patents. SerpKite integrates natively with popular AI frameworks like LangChain, CrewAI, LlamaIndex, and provides an MCP server for Claude, Cursor, and ChatGPT. The platform uses a credit-based pricing model starting at $0.60 per 1,000 searches, with credits that never expire and no charges for failed requests. It offers official SDKs for TypeScript, Python, and Go, plus batch processing at 50% discount and caching to optimize costs.
Key features
LLM-optimized output formats
Returns results as clean JSON or Markdown with token-efficient compact mode, reducing token usage by up to 66% compared to standard JSON.
Complete SERP features in one call
Fetches organic results, knowledge graph, answer box, People Also Ask, related searches, and sitelinks all in a single API request.
Official multi-language SDKs
Typed clients for TypeScript, Python, and Go with built-in retries, timeouts, and batch polling.
Native AI framework integrations
Direct support for LangChain, CrewAI, LlamaIndex, and MCP servers for Claude, Cursor, and ChatGPT.
Multiple search verticals
Access Google Search, News, Images, Videos, Maps, Places, Reviews, Shopping, Scholar, Patents, and Autocomplete from one API.
Batch processing at discount
Queue up to 100 queries per call at 50% off standard rates with webhook delivery of results.
Use cases
- 1
AI agents and autonomous systems
Provide real-time web search capabilities to AI agents and LLMs for research, fact-checking, and decision-making.
- 2
LLM-powered applications
Integrate live search results into chatbots, assistants, and language model applications with minimal token overhead.
- 3
Keyword research and SEO tools
Use autocomplete and search APIs to build keyword research, rank tracking, and SEO analysis tools.
- 4
No-code automation workflows
Connect search capabilities to n8n, Make, Zapier, and other automation platforms for data enrichment and monitoring.
FAQ
What is a credit?
One credit is one Google results page (up to 10 organic results) from /v1/search or any core vertical. Autocomplete, batch jobs, and cache hits cost 0.5 credits.
Do credits expire?
No. Purchased credits stay on your balance until you spend them, and credit packs stack with no expiration.
Am I charged for failed requests?
No. SerpKite refunds upstream failures, empty results, and blocked requests. Every billed response includes X-Credits-Used and X-Credits-Remaining headers.
What SDKs are available?
Official SDKs for TypeScript (npm install serpkite), Python (pip install serpkite), and Go, plus langchain-serpkite for LangChain and serpkite[crewai] for CrewAI.
How does the MCP server work?
SerpKite offers a remote MCP server at api.serpkite.com/v1/mcp that integrates with Claude, Cursor, ChatGPT, and any MCP-compatible client for native search capabilities.
Tech stack & tags
Feedback & Discussion
Discussion
Sign in to leave a comment or rating.
No comments yet. Be the first.