Ohu Logo OHU
Home Use Cases Pricing FAQs Contact MCP Server Docs Dashboard
Streamable HTTP Transport (MCP V1.0)

Empower Your AI Agents with Ohu Web Intelligence

Ohu provides a native Model Context Protocol (MCP) server over Streamable HTTP. Connect OpenCode CLI, Claude Desktop, Cursor, Windsurf, or autonomous LLM agents to 17 high-performance tools for web scraping, RAG chunking, brand sentiment monitoring, AEO auditing, and security scanning.

Remote MCP Streamable Endpoint https://sluxia.com/ohu/mcp
Smithery.ai Registry
Standard Remote MCP Configuration Specification:
{ "mcpServers": { "ohu": { "url": "https://sluxia.com/ohu/mcp", "headers": { "Authorization": "Bearer YOUR_OHU_API_KEY", "Accept": "application/json, text/event-stream" } } } }

Harness-Specific Client Guides

Select your AI workspace to view its specific JSON configuration schema:

OpenCode CLI Setup

Add to ~/.config/opencode/opencode.jsonc (requires explicit "type": "remote" and "enabled": true):

{ "mcp": { "ohu": { "type": "remote", "url": "https://sluxia.com/ohu/mcp", "enabled": true, "headers": { "Authorization": "Bearer YOUR_OHU_API_KEY", "Accept": "application/json, text/event-stream" } } } }

Claude Desktop Setup

Add to claude_desktop_config.json under mcpServers:

{ "mcpServers": { "ohu": { "url": "https://sluxia.com/ohu/mcp", "headers": { "Authorization": "Bearer YOUR_OHU_API_KEY", "Accept": "application/json, text/event-stream" } } } }

Cursor IDE & Windsurf Setup

Add to .cursor/mcp.json or Cursor Settings > Features > MCP:

{ "mcpServers": { "ohu": { "url": "https://sluxia.com/ohu/mcp", "headers": { "Authorization": "Bearer YOUR_OHU_API_KEY", "Accept": "application/json, text/event-stream" } } } }

Continue.dev & Zed Setup

Add to ~/.continue/config.json under experimental.modelContextProtocol:

{ "name": "ohu", "type": "sse", "url": "https://sluxia.com/ohu/mcp", "requestOptions": { "headers": { "Authorization": "Bearer YOUR_OHU_API_KEY", "Accept": "application/json, text/event-stream" } } }

17 Available MCP Agent Tools

Registered in mcp/tools.js
brand_mentions Intelligence

Monitors online brand mentions across web & news while excluding self domain, analyzing sentiment breakdown and market share of voice.

params: brand_name, brand_domain, keywords, recency, max_results
search Web

Performs live web search across multiple search engines with domain authority ranking and proximity snippet highlights.

params: query, max_results, region, time_range, category
aeo Optimization

Analyzes webpage indexability and citation probability for AI search engines, outputting readiness scores and schema snippets.

params: url
rag_chunk Ingestion

Converts webpages or raw text into optimized chunks for vector databases, preserving heading hierarchy and generating Pinecone formats.

params: url, text, chunk_size, overlap
security_audit Security

Scans webpages for security vulnerabilities, hidden prompt injection risks targeting AI models, and leaked API secret keys.

params: url, html
synthetic_dataset AI Training

Generates high-quality question and answer instruction pairs from web documentation for fine-tuning LLM models.

params: url, html
crawl Scraping

Crawls website pages and returns content in multiple formats including Markdown, raw HTML, plain text, and extracted links.

params: url, formats, css_selector, max_length, max_pages
markdown Conversion

Converts web pages or HTML elements into clean Markdown format, expanding relative links and retaining images.

params: url, css_selector, max_length
geo_competitor Optimization

Compares your website against a competitor for AI search visibility, analyzing semantic relevance and content gaps.

params: your_url, competitor_url, topic
llms_generator Manifest

Generates standard /llms.txt and /llms-full.txt markdown manifests for your website to help AI models navigate your docs.

params: url, company_name, description
pdf_to_markdown Parser

Converts PDF documents into clean, structured Markdown text while preserving multi-column layouts and headings.

params: url, max_length
table_extract Extraction

Extracts tables from any webpage, automatically inferring data types and exporting formatted JSON, Markdown tables, and CSV.

params: url, html
clean_article Distillation

Extracts clean article content from any webpage by removing ads and navigation menus—returning text, summary, and hero images.

params: url, html, max_length
code_extract Extraction

Extracts code blocks from technical docs, identifying languages and separating code snippets from terminal commands.

params: url, html
media_extractor Assets

Extracts all media assets from a webpage, including video streams, audio podcasts, high-res images, and hero images.

params: url, html
link_graph Graph

Extracts and categorizes all links on a webpage into internal pages, external sites, and social profiles.

params: url, html
extract Scraping

Extracts structured data from any webpage using CSS selectors, supporting attribute targets and nested schemas.

params: url, selectors

OAuth 2.1 & Specification Endpoints

Ohu implements official OAuth 2.1 specifications (RFC 9728, RFC 7662, RFC 7591) for automatic AI agent discovery:

Protected Resource Metadata
GET /.well-known/oauth-protected-resource
Token Introspection
POST /ohu/api/v1/oauth/introspect
Dynamic Client Registration
POST /ohu/api/v1/oauth/register