Overview
What WebPulse does
WebPulse gives AI agents two ways to work with the web.
- POST /extract with {"url": "https://example.com"} fetches the page and returns its title, meta description and a markdown version of the content (headings, paragraphs, lists and links; scripts, styles, navigation, headers and footers removed), capped at 8,000 characters, with its length.
- POST /search with {"query": "best practices for serverless", "num": 5} returns up to 10 results (default 5), each with title, URL, snippet and, when the source has one, its published date. Queries can be up to 200 characters.
Each result carries its source URL and title. Display the source link with any output that uses a result. Results must not be stored or extracted in bulk.
Where the data comes from
- /extract reads the page you name, live, with a 10-second timeout; a page that cannot be fetched answers 502 and is not billed.
- /search runs on AWS Web Search on Amazon Bedrock AgentCore. Search results are not cached. If the search service is unavailable the API answers 502 or 503 and the call is not billed.
Three ways to connect
- REST with an X-API-Key header at https://webpulse.waltsoft.net (routes at the root, no /v1), described by an OpenAPI 3 schema (https://webpulse.waltsoft.net/openapi.json ).
- MCP: the tool webpulse___extractWebContent (url, api_key) on the Amazon Bedrock AgentCore Gateway.
- A2A (Agent2Agent 0.3): the agent card at https://webpulse.waltsoft.net/.well-known/agent-card.json with the skills extractContent and searchWeb.
Who it is for
AI agents that need to read a web page or find current sources without running a browser or scraper, and research and RAG pipelines that need page text as markdown with its source.
Billing
On AWS Marketplace each successful call is one unit at $0.01, except POST /search, which uses 2 units ($0.02). Errors (400, 401, 404, 429, 5xx) are never billed. With x402 (USDC on Base or Solana, no account): /extract $0.01 and /search $0.02 per call. Rate limit: 100 requests per second (bursts up to 50).
Highlights
- Web page to clean markdown in one call: POST a URL to /extract and get the title, meta description and markdown content (scripts, navigation, headers and footers removed), up to 8,000 characters
- Web search on AWS Web Search on Amazon Bedrock AgentCore: up to 10 results per query, each with its source URL, title and snippet; one $0.02 search uses 2 Marketplace units
- Agent-ready: REST with an OpenAPI 3 schema, an MCP tool on Amazon Bedrock AgentCore Gateway and an A2A agent card with extractContent and searchWeb; only successful calls are billed
Details
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
Quick Launch
Pricing
Dimension | Description | Cost/unit |
|---|---|---|
API Call | One unit per extract; a web search (POST /search) uses 2 units | $0.01 |
Vendor refund policy
Contact support@waltsoft.net within 30 days for refund requests.
How can we make this page better?
Legal
Vendor terms and conditions
Content disclaimer
Delivery details
- Amazon Bedrock AgentCore
API-Based Agents & Tools
API-Based Agents and Tools integrate through standard web protocols. Your applications can make API calls to access agent capabilities and receive responses.
Additional details
Usage instructions
To get started using the remote MCP server, follow the instructions below:
Available Tools This MCP server (Amazon Bedrock AgentCore Gateway) supports the following tool:
- webpulse___extractWebContent - Extracts a web page as clean markdown (title, description, content capped at 8,000 characters, length). Arguments: url (required), api_key (required). Web search (POST /search) is available over REST and A2A, not as an MCP tool.
Prerequisites
- An active subscription to this product. Your API key (it starts with mp_) is stored in your AWS Secrets Manager after you choose Set up your account, and is also shown on the registration page.
- Install Node.js and npm (for mcp-remote), or Python 3.10+ (for the Strands Agents SDK).
Authentication The MCP endpoint needs no connection header. Pass your key as the api_key argument of every tool call; replace YOUR_API_KEY with your actual key below. A wrong or inactive key returns status 401.
MCP endpoint: https://waltsoft-marketplace-gateway-xncyjgs1uh.gateway.bedrock-agentcore.us-east-1.amazonaws.com/mcp
Claude Desktop Edit the configuration file at:
- macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
- Windows: %APPDATA%\Claude\claude_desktop_config.json
Add the below code:
{ "mcpServers": { "waltsoft-webpulse": { "command": "npx", "args": [ "mcp-remote", "https://waltsoft-marketplace-gateway-xncyjgs1uh.gateway.bedrock-agentcore.us-east-1.amazonaws.com/mcp" ] } } }Then ask, for example: "Use webpulse___extractWebContent with url https://example.com and api_key YOUR_API_KEY".
Strands Agents (Python)
pip install strands-agents mcp from mcp.client.streamable_http import streamable_http_client from strands.tools.mcp import MCPClient URL = "https://waltsoft-marketplace-gateway-xncyjgs1uh.gateway.bedrock-agentcore.us-east-1.amazonaws.com/mcp" with MCPClient(lambda: streamable_http_client(URL)) as mcp: result = mcp.call_tool_sync("call-1", "webpulse___extractWebContent", {"url": "https://example.com", "api_key": "YOUR_API_KEY"}) print(result["content"][0]["text"])Tool result shape The tool returns {"statusCode": "integer", "body": "string (JSON)", "result": {...}}; result is the POST /extract response:
{"url": "string", "title": "string", "description": "string", "content": "string (markdown)", "length": "integer", "cached": "boolean"}Billing Only successful calls are billed: one $0.01 unit per extraction.
Rate Limits
- 100 requests per second (bursts up to 50).
- Exceeding returns HTTP 429 Too Many Requests.
- Use retry and exponential backoff to handle limits.
Learn More API Docs: https://webpulse.waltsoft.net Support: support@waltsoft.net
Support
Vendor support
Email support@waltsoft.net for API questions and integration help. Response within 24 hours.
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.