Correctover MCP Server - Contract Validation & API Failover¶

Repo: Correctover/mcp-server Install: npx -y correctover-mcp-server Pricing: Open source Category: Legal / API Infrastructure

Overview¶

Correctover provides two critical capabilities for business operators who rely on AI: 1. Contract validation - 6-dimension verification (structure, schema, latency, cost, identity, integrity) at 22μs P50 2. Self-healing LLM API failover - 87 self-healing rules with MAPE-K autonomic loop across 9 LLM providers

For operators running AI agents in production, this means contracts are validated before execution and LLM API calls automatically fail over when a provider goes down - no manual intervention needed.

Why Business Operators Need This¶

Contract Risk Mitigation - Before your agent signs or executes any contract, Correctover validates it across 6 dimensions. Catches errors, missing clauses, and structural issues in microseconds.
API Reliability - If your primary LLM provider (OpenAI, Anthropic, etc.) has an outage, Correctover automatically fails over to an alternative provider in under 1 second (949ms E2E). Your agents keep running.
Cost Control - The cost dimension validation ensures you're not overpaying for API calls. BYOK (Bring Your Own Key) to 9 providers for direct, transparent pricing.

Installation¶

{
  "mcpServers": {
    "correctover": {
      "command": "npx",
      "args": ["-y", "correctover-mcp-server"]
    }
  }
}

Supported LLM Providers (9)¶

Provider	Type	Failover Priority
OpenAI	Primary	L1
Anthropic	Primary	L1
DeepSeek	Secondary	L2
Moonshot	Secondary	L2
Zhipu AI	Secondary	L2
Qwen	Secondary	L2
SiliconFlow	Tertiary	L3
Groq	Tertiary	L3
Together AI	Tertiary	L3

Key Tools¶

Tool	Function	Latency
`validate_contract`	6-dimension contract verification	22μs P50
`health_check`	Provider health monitoring	Real-time
`failover_status`	Current failover state and routing	Instant
`cost_estimate`	Token cost estimation across providers	< 100ms
`schema_check`	Contract schema validation	Microseconds
`identity_verify`	Contract signer identity verification	Varies

Common Queries¶

Contract Validation¶

"Validate this contract before I sign it. Check structure, schema, and identity."
"Run a full 6-dimension verification on the vendor agreement."
"Check this NDA for missing standard clauses."

API Failover Management¶

"Show current LLM provider health status. Which providers are available?"
"Set up failover: primary OpenAI, fallback to Anthropic, tertiary to DeepSeek."
"What was the failover event from last night? Which provider went down?"

Cost Optimization¶

"Compare token costs across all 9 providers for a 10K input / 2K output query."
"Which provider is cheapest for GPT-4-level quality right now?"
"Alert me if any provider raises prices by more than 10%."

Best Practices¶

Always validate contracts before execution - 22μs is faster than human review by 6 orders of magnitude
Configure failover chains - Set up L1 (OpenAI/Anthropic) → L2 (DeepSeek/Moonshot) → L3 (Groq/Together) for maximum reliability
Monitor the health dashboard - Check provider health before starting long agent workflows
Use cost_estimate proactively - Compare provider pricing weekly; prices change frequently
BYOK for each provider - Direct API keys give you the best pricing and avoid middleware markups

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects

Limitations¶

New server (June 2026) - production track record still developing
Contract validation focused on structure and schema, not legal advice (use a lawyer for legal review)
Failover latency (949ms) may be noticeable in real-time chat applications
Requires API keys for each provider you want to fail over to

For Operators: Integration with CorpusIQ¶

CorpusIQ Stripe connector + Correctover = Validate payment contracts and monitor API costs
CorpusIQ QuickBooks connector + Correctover = Track LLM API spending alongside other operational costs
CorpusIQ cross-source analysis + Correctover = Correlate API cost data with business outcomes