Orthogonal MCP Server認証済み注目

Connect your AI agent to paid tools and data through one MCP server.

Orthogonal MCP gives Claude, Cursor, ChatGPT, Codex, and other MCP-compatible clients access to a catalog of verified APIs without requiring a separate integration for every provider. Your agent can discover the right tool, call it, and pay per request through Orthogonal.

Why use Orthogonal MCP?

AI agents are useful when they can access current information and take action through external services. However, connecting an agent to every API usually means managing separate accounts, API keys, authentication flows, contracts, billing systems, and response formats.

Orthogonal provides a unified access layer:

• One MCP connection for many tools and data providers
• Searchable access to company data, contact information, web research, scraping, financial data, and more
• A consistent way for agents to discover and call APIs
• Per-request pricing instead of separate subscriptions for every provider
• Managed provider connections and consolidated billing
• A catalog of verified and monitored services
• Support for MCP clients such as Claude and Cursor

Orthogonal’s catalog includes hundreds of endpoints from dozens of providers. The available tools, prices, and authentication requirements can change as the catalog grows, so agents should search the catalog at runtime rather than relying on a fixed list of integrations.

How Orthogonal MCP works

Orthogonal MCP follows a simple three-step workflow:

1. Discover — The agent describes what it needs, such as “find technology companies in Toronto with more than 100 employees.”
2. Call — Orthogonal selects or exposes the appropriate API and sends the request to the provider.
3. Pay — The request is metered and charged according to the selected tool. The agent receives the result along with usage information.

This means an agent can move from an intent to a live result without a developer manually wiring together every provider.

Connect your MCP client

Add Orthogonal as an MCP server in a compatible client:

{
  "mcpServers": {
    "orthogonal": {
      "url": "https://mcp.orthogonal.com"
    }
  }
}

After connecting, authenticate using your Orthogonal account or API credentials according to the current setup instructions. Once configured, the same MCP connection can be used to search for available tools and run supported APIs.

For setup instructions and client-specific configuration, see the Orthogonal documentation.

Example prompts

Find a tool that can search the live web and return source URLs.

Find a company-data API that can identify software companies in Canada with more than 50 employees.

Find a tool that can extract structured product information from a public webpage.

Find the cheapest verified API that can enrich a business record with its website, industry, location, and employee count.

The agent can search for the appropriate service, report the expected cost, run the tool, and return the result in the same conversation.

API-backed use cases and workflows

The examples below are based on API categories currently available through the Orthogonal catalog. Exact availability, pricing, limits, and required fields may vary by provider.

Company discovery and enrichment

Relevant APIs include FullEnrich, Bytemine, Company Enrich, People Data Labs, Crustdata, Abstract, Coresignal, and Tomba.

Find companies that match a target profile

Find 100 B2B SaaS companies in the United States that:

- Have between 50 and 500 employees
- Sell software to marketing teams
- Have raised funding in the last five years
- Are currently hiring
- Have an identifiable head of sales or VP of marketing

Return the company name, website, location, employee range, funding information, hiring signal, and source URLs.

Enrich a company list

Enrich the following company domains:

- example-one.com
- example-two.com
- example-three.com

Return:

- Company name
- Website
- Industry
- Headquarters
- Employee count
- Estimated revenue range
- Founding year
- Funding stage
- Technology stack
- LinkedIn URL
- Confidence notes

Build a target-account list

Create a target-account list for a company selling developer infrastructure.

Find companies that:

- Have at least 100 employees
- Employ software engineers
- Use cloud infrastructure
- Are headquartered in Canada, the United States, or the United Kingdom
- Have recently hired engineering or platform roles

Score each account from 1 to 5 and explain the score.

Contact discovery and enrichment

Relevant APIs include FullEnrich, Bytemine, Company Enrich, Crustdata, People Data Labs, Findymail, Tomba, Icypeas, and Fiber AI.

Find decision-makers at target accounts

For each company in this list, find the most relevant decision-maker for an enterprise security product.

Prioritize:

- Chief Information Security Officer
- VP of Security
- Head of Infrastructure
- VP of Engineering

Return the person’s name, title, company, public profile, location, and evidence that they currently hold the role.

Find and verify professional email addresses

Find the professional email address for each person below.

For every result:

- Return the email address
- Identify the associated company
- Verify deliverability
- Include the confidence level
- Do not return personal email addresses
- Mark unverifiable results as unknown

Reverse lookup an email address

For each email address below, identify the associated person and company.

Return:

- Full name
- Job title
- Company
- Company website
- Public professional profile
- Confidence level
- Whether the address appears deliverable

Enrich a CRM export

Enrich this CSV of contacts with:

- Current job title
- Current company
- Company website
- LinkedIn URL
- Work email
- Email verification status
- Company employee count
- Industry

Preserve the original columns and add the new fields. Do not overwrite existing values unless the new value is more recent and verified.

Web search and research

Relevant APIs include Massive Web API, Context.dev, Serper, Seltz, ScrapeGraphAI, and Tavily.

Research a market

Research the market for AI-powered sales platforms.

Find:

- Major vendors
- Emerging startups
- Target customers
- Pricing models
- Key product categories
- Recent funding
- Recent acquisitions
- Main differentiators
- Important risks

Use current web sources, include links for every important claim, and separate verified facts from analysis.

Create a daily news briefing

Create a daily briefing about the cybersecurity industry.

Include:

- Five important news stories
- Funding announcements
- Acquisitions
- Major breaches
- Regulatory developments
- Product launches

For each item, provide the publication date, company, two-sentence summary, and source URL.

Fact-check a report

Fact-check the following report using current public sources.

For each claim, return:

- Supported
- Contradicted
- Unclear
- Evidence
- Source URL
- Publication date
- Suggested correction

Do not treat search snippets as sufficient evidence.

Web scraping and structured extraction

Relevant APIs include ScrapeGraphAI, Bytemine, Context.dev, Serper Scrape, Olostep, Notte, and Tavily Extract.

Extract data from a webpage

Extract all products from this webpage.

Return JSON with:

- Product name
- Description
- Price
- Currency
- Availability
- Product URL
- Rating
- Review count

Use null when a field is missing. Do not infer values.

Crawl a website

Crawl this documentation site and create a structured index.

For every relevant page, return:

- Page title
- URL
- Main topic
- Product area
- Key concepts
- Referenced API endpoints
- Last-updated date if available

Analyze a competitor’s website

Analyze this competitor’s website.

Identify:

- Product positioning
- Target customer
- Main use cases
- Pricing
- Integrations
- Customer proof
- Calls to action
- Technical claims
- Pages that appear recently updated

Return a structured report with source URLs.

Batch-process URLs

Process the following 50 URLs in parallel.

For each URL:

- Extract the company name
- Identify the product category
- Summarize the page in 100 words
- Detect pricing information
- Detect contact information
- Return the page title and canonical URL

Return one structured record per URL and report failed pages separately.

Local-business discovery

Relevant APIs include Openmart, Fiber AI, Serper Places, Serper Maps, Serper Reviews, Voygr, and Tomba location data.

Find local businesses

Find independent dental clinics in Toronto that:

- Have fewer than five locations
- Have a working website
- Have at least 20 public reviews
- Are open on Saturdays
- Do not appear to use online appointment scheduling

Return the business name, address, phone number, website, rating, review count, and source.

Build a local-business lead list

Find 200 restaurants in Vancouver that offer delivery and have fewer than 50 employees.

Return:

- Restaurant name
- Address
- Phone
- Website
- Cuisine
- Rating
- Review count
- Delivery availability
- Owner or decision-maker when publicly available

Analyze local-business reviews

Analyze the latest public reviews for these five businesses.

Identify:

- Most common complaints
- Most praised features
- Service issues
- Repeated product requests
- Changes in sentiment over time

Include representative source links and distinguish review evidence from interpretation.

Financial and company intelligence

Relevant APIs include Financial Datasets, Company Enrich, People Data Labs, Coresignal, and Aviato.

Create a public-company briefing

Create a current briefing for these public companies:

- Company A
- Company B
- Company C

Include:

- Latest stock price
- Market capitalization
- Revenue
- Revenue growth
- Profitability
- Recent earnings
- SEC filings
- Recent company news
- Key risks

Include the data timestamp and source for every financial figure.

Compare companies

Compare these three public companies.

Create a table covering:

- Revenue
- Revenue growth
- Gross margin
- Operating margin
- Free cash flow
- Valuation multiples
- Recent earnings trend
- Major risks

Use the latest available data and identify any values that are not directly comparable.

Build a market map

Create a market map for companies similar to this business:

Company: example.com

Group companies by:

- Customer segment
- Product category
- Company size
- Geography
- Funding stage
- Business model

For each company, include the website and a short explanation of why it belongs in the category.

Social and creator research

Relevant APIs include Scrape Creators, Serper, Massive Web API, Seltz, and Context.dev.

Find social content by topic

Find recent public TikTok and YouTube content about developer tools.

Return:

- Creator
- Platform
- Video title
- URL
- Publication date
- View count if available
- Engagement signals
- Main topic
- Sentiment

Group the results by recurring theme.

Analyze a creator or channel

Analyze this creator’s recent content.

Identify:

- Main topics
- Audience
- Posting frequency
- Highest-performing themes
- Repeated products or brands
- Common audience questions
- Potential partnership fit

Use public information only and include the relevant content URLs.

Recruiting and workforce intelligence

Relevant APIs include Company Enrich, People Data Labs, Crustdata, Coresignal, FullEnrich, and web-search or scraping APIs.

Analyze hiring trends

Analyze the hiring activity of these companies over the last 90 days.

Return:

- Number of open roles
- Engineering roles
- Sales roles
- Marketing roles
- Executive roles
- Locations
- Remote or hybrid signals
- Repeated skills
- New functions or departments

Use official careers pages where possible.

Find companies hiring for a skill

Find companies in the United States that are currently hiring for machine-learning infrastructure roles.

Prioritize companies that:

- Have at least 50 employees
- Have raised funding or report meaningful revenue
- Have more than one relevant open role
- Have an engineering team in North America

Return the company, role, location, job URL, and hiring signal.

Technology-stack research

Relevant APIs include Bytemine, ScrapeGraphAI, Company Enrich, Abstract, Coresignal, and web-search APIs.

Detect technologies used by a website

Analyze this website and identify likely technologies used for:

- Frontend
- Backend
- Hosting
- Analytics
- Payments
- Customer support
- Marketing automation
- Content management
- Authentication

Separate direct evidence from inference and include the evidence URL.

Find prospects using a technology

Find companies in the United States that appear to use HubSpot and have between 20 and 200 employees.

Return:

- Company name
- Website
- Industry
- Employee range
- Evidence of HubSpot usage
- Relevant decision-maker
- Source URL

Multi-API agent workflows

The strongest workflows combine multiple Orthogonal APIs in sequence.

Research, enrich, and prioritize accounts

Build a prioritized account list for our sales team.

1. Find companies matching our target customer profile.
2. Enrich each company with industry, employee count, funding, location, and technology data.
3. Find relevant executives and verify their work emails.
4. Search for recent company news and hiring signals.
5. Score each account based on fit and urgency.
6. Return a CSV-ready table with source URLs, confidence notes, and estimated tool usage.

Possible API categories:

• Company search
• Company enrichment
• Contact enrichment
• Email finding
• Email verification
• Web search
• News search
• Website scraping

Find local businesses and identify owners

Find independent fitness studios in Chicago with strong reviews but outdated websites.

1. Search local businesses.
2. Retrieve ratings, reviews, addresses, and websites.
3. Crawl each website.
4. Identify signs of outdated technology or missing conversion features.
5. Find the owner or relevant decision-maker.
6. Verify a professional contact method.
7. Return a prioritized outreach list with evidence.

Build a competitor-monitoring report

Create a weekly competitor report.

For each competitor:

1. Crawl the pricing and product pages.
2. Detect changes from the previous report.
3. Search for recent announcements and news.
4. Track hiring activity.
5. Identify new integrations or positioning changes.
6. Summarize the strategic implications.
7. Include evidence links for every change.

Create a source-backed investment brief

Create a research brief for this public company.

1. Retrieve recent financial metrics.
2. Retrieve recent filings.
3. Search for company and industry news.
4. Review the company website and investor materials.
5. Compare the company with selected peers.
6. Summarize opportunities, risks, and unanswered questions.
7. Include timestamps and sources for every important figure.

Create a structured research dataset

Create a structured dataset of 1,000 companies in the climate-tech sector.

For each company, collect:

- Name
- Website
- Location
- Industry
- Employee range
- Funding stage
- Founders
- Key executives
- Public email
- Technology signals
- Recent news
- Source URLs

Deduplicate records, validate domains, flag uncertain fields, and return failed or incomplete records separately.

Reliability and control patterns

Ask for sources

Include a source URL for every externally verifiable claim.

Ask the agent not to guess

Do not infer missing values. Return “unknown” and explain what could not be verified.

Request structured output

Return JSON with the fields:
name, website, category, evidence, source_url, confidence.

Use official sources where possible

Use the company’s official documentation or website as the primary source, then use secondary sources only to fill gaps.

Require approval before expensive workflows

Before running this workflow:

1. List the APIs you plan to use.
2. Explain what each API will retrieve.
3. Estimate the number of calls.
4. Provide an estimated cost range.
5. Identify any sensitive data involved.
6. Wait for my confirmation.

Set a budget

Complete this workflow with a maximum tool budget of $5.

Prefer the lowest-cost verified API that meets the requirements.
Use batch endpoints when available.
Stop and ask for confirmation if the estimated cost exceeds the budget.
Report the APIs used, number of calls, and final usage.

Best practices

• Use discovery instead of hard-coding provider names whenever possible.
• Ask for source URLs and timestamps for research results.
• Treat retrieved data and agent-generated analysis as separate fields.
• Set limits for the number of records, pages, or requests.
• Review provider-specific pricing before running large workflows.
• Avoid sending sensitive information unless the selected provider is appropriate.
• Require confirmation before taking actions that change external systems.
• Validate important results against primary sources.
• Record the tool used, request parameters, cost, and timestamp for repeatability.

The result

With Orthogonal MCP, an agent can move from a natural-language request to a complete research or data workflow:

Discover → Enrich → Search → Extract → Verify → Structure → Report

One MCP connection can support workflows across company intelligence, contact data, live web research, scraping, financial analysis, local-business discovery, social research, and structured data collection.

Get started

Connect Orthogonal MCP to your preferred MCP-compatible client, then ask your agent to find a tool for the task you need to complete.

Find the best available tool for this task, explain the expected cost, and wait for my confirmation before running it.

Once connected, Orthogonal lets your agent discover, call, and pay for external tools through a single interface.

• Visit Orthogonal
• Read the MCP documentation
• Browse the API catalog