Simple, transparent pricing
Choose the plan that fits your needs. Scale as you grow.
Compare plans
Find the perfect plan for your needs. All plans include core MCP infrastructure.
Pilot
Infrastructure
- MCP servers: Up to 3
- Included MCP requests: 500K / month
- Overage pricing: $0.30 per 1K
- Spend cap & overage alerts
- Preview & production environments
- Promote preview to production
- Drift detection: Basic
MCP Generation
- OpenAPI to MCP conversion
- Regenerate on spec updates
- Automatic spec optimization
- Improved tool naming and schemas
Intelligence
- Tool usage analytics & agent behavior insights
Security & Access
- OAuth and API key pass-through
- Team members: Unlimited
- Roles & permissions: Admin only
- Tool-level permissions
Analytics & Logs
- Usage metering and attribution
- Log retention: 7 days
Support
- Support level: 48hr email
- SLA: 99.5%
Core
Infrastructure
- MCP servers: Up to 10
- Included MCP requests: 2M / month
- Overage pricing: $0.25 per 1K
- Spend cap & overage alerts
- Preview & production environments
- Promote preview to production
- Drift detection: Basic
- Custom domains
MCP Generation
- OpenAPI to MCP conversion
- Auto-generated MCP documentation
- Regenerate on spec updates
- Automatic spec optimization
- Improved tool naming and schemas
Intelligence
- Tool usage analytics & agent behavior insights
Security & Access
- OAuth and API key pass-through
- Team members: Unlimited
- Roles & permissions: Admin & Viewer
- Tool-level permissions
- Advanced access controls
Analytics & Logs
- Usage metering and attribution
- Usage export
- Log retention: 14 days
- Log export to your storage
Support
- Support level: Priority email
- SLA: 99.9%
Growth
Infrastructure
- MCP servers: Up to 25
- Included MCP requests: 10M / month
- Overage pricing: $0.20 per 1K
- Spend cap & overage alerts
- Preview & production environments
- Promote preview to production
- Drift detection
- Custom domains
MCP Generation
- OpenAPI to MCP conversion
- Auto-generated MCP documentation
- Regenerate on spec updates
- Automatic spec optimization
- Improved tool naming and schemas
Intelligence
- Tool usage analytics & agent behavior insights
- Configurable intelligence policies
Security & Access
- OAuth and API key pass-through
- Team members: Unlimited
- Roles & permissions: Admin, Developer & Viewer
- Tool-level permissions
- Advanced access controls
Analytics & Logs
- Usage metering and attribution
- Usage export
- Audit logging
- Log retention: 30 days
- Log export to your storage
Support
- Support level: Priority + onboarding
- SLA: 99.9%
Enterprise
Infrastructure
- MCP servers: Unlimited
- Included MCP requests: 25M+ / month
- Overage pricing: Volume-based
- Spend cap & overage alerts
- Preview & production environments
- Promote preview to production
- Drift detection
- Custom domains
- Tenant isolation
MCP Generation
- OpenAPI to MCP conversion
- Auto-generated MCP documentation
- Regenerate on spec updates
- Automatic spec optimization
- Improved tool naming and schemas
Intelligence
- Tool usage analytics & agent behavior insights
- Configurable intelligence policies
Security & Access
- OAuth and API key pass-through
- Team members: Unlimited
- Roles & permissions: Custom roles + IdP sync
- Tool-level permissions
- Advanced access controls
- SSO and SCIM
Analytics & Logs
- Usage metering and attribution
- Usage export
- Audit logging
- Log retention: Configurable
- Log export to your storage
Support
- Support level: Dedicated
- SLA: Custom (financially backed)
Frequently asked questions
Everything you need to know about Trevi, MCP servers, and how we bridge agents to your existing APIs.
MCP (Model Context Protocol) servers let AI agents interact with your APIs in a standardized way. Instead of building custom integrations for each agent, MCP provides a common interface. Trevi generates and hosts these servers from your existing API specs, so you don't need to build them yourself.
No. Trevi works with your existing API as-is. We read your OpenAPI or Postman spec and generate an MCP server that acts as a bridge between agents and your API. Your API stays unchanged.
Each MCP server runs in a secure, production-grade environment. You configure API keys and authentication that Trevi uses when calling your API. We also validate all agent inputs before forwarding requests to prevent malformed calls.
Yes. You can generate and test MCP servers with staging API endpoints before deploying to production. Each server is contained, so you can have separate staging and production instances.
When you update your spec, Trevi can regenerate the MCP server with the new changes. You can preview the changes before deploying, then redeploy with one click to update your live endpoint.
Trevi supports OpenAPI 3.0+ (YAML or JSON) and Postman collections. We validate and normalize your spec to ensure compatibility with MCP generation.