Veroq MCP

What veroq-ai/veroq-mcp MCP server does

The veroq-ai/veroq-mcp MCP server gives MCP-compatible agents access to financial research and verification workflows through 62 tools. Its finance features include price data, technical analysis, earnings, news, sentiment, market movers, economic data, foreign exchange, commodities, sectors, screening, alerts, crypto, insider activity, SEC filings, analyst information, congressional activity, and institutional holdings.

The primary entry points are veroq_ask and veroq_verify. The first routes a financial question across 41 intents, while the second checks a claim and returns an evidence chain with confidence information. More specialized tools analyze individual tickers, compare securities, verify market claims, produce composite trading signals, and generate broader market briefings.

Verification also extends to multi-agent workflows. veroq_run_verified_swarm runs a five-stage process involving planning, research, verification, criticism, and synthesis. veroq_create_runtime creates a domain-specific workflow for finance, legal, research, compliance, or custom use cases. The package also includes feedback processing, report generation and retrieval, agent execution, memory operations in the programmatic library, and a tool for calling registered external MCP servers.

How it works

MCP clients invoke named VeroQ tools and receive results from the VeroQ service. Verification-oriented responses can include evidence chains, confidence breakdowns, and source reliability scores. The verified swarm checks work at each stage and can escalate high-stakes cases according to the service configuration described in the project documentation.

Workflows can use cost modes named cheap, balanced, and premium, with per-step budgets and credit reporting. The documentation also describes permission rules, decision lineage, human review escalation, audit trails, external MCP trust levels, rate limits, feedback handling, and observability features for enterprise customers.

Setup and configuration

Install the veroq-ai/veroq-mcp MCP server globally with npm:

npm install -g veroq-mcp

Create a VeroQ API key through the VeroQ settings page, then provide it as VEROQ_API_KEY. The optional VEROQ_BASE_URL variable changes the API base URL; the documented default is https://api.veroq.ai. POLARIS_API_KEY and POLARIS_BASE_URL are accepted for backward compatibility.

For Claude Desktop, configure the command as veroq-mcp and pass VEROQ_API_KEY in the server environment. Cursor uses the same command and environment variable through its MCP server settings. The package is also available as a TypeScript/Python library for creating runtimes, registering external MCP servers, verifying arbitrary LLM output, storing and recalling agent memory, and subscribing to verification events.

Tools and capabilities

The documented tool groups include:

• General questions and claim verification
• Ticker analysis, market data, screening, alerts, and signals
• Crypto, DeFi, search, web extraction, research, forecasts, and timelines
• SEC filings, insider trades, analysts, congressional activity, and institutions
• Verified swarms, domain runtimes, reports, feedback, and external MCP calls

Many operations consume credits. The listed costs range from free tool discovery and feedback operations to 1–5 credits for general questions, 3 credits for several analysis tools, and 15–25 credits for verified swarms. Exact availability and pricing should be checked against the VeroQ service.

Limitations and notes

The server requires a VeroQ API key, and the repository points users to VeroQ pricing for service details. It is primarily oriented toward financial intelligence, even though its runtime supports other domains. External MCP calls require registered servers and are subject to the permission, trust, and rate controls provided by VeroQ. Trading signals and verification results are outputs for agent workflows; the supplied material does not describe them as investment advice or guarantee their accuracy.