Service Discovery
List all services in your mesh with their versions, endpoints, health status, and traffic weights. Get a real-time topology view of your entire service graph.
MCP Server · npm package
AI Microservices Orchestration via Service Mesh
An MCP server that gives AI agents a unified interface for your entire service mesh. Service discovery, traffic routing, health monitoring, circuit breakers, and multi-service workflow orchestration — all programmable from any AI agent.
$49/mo. Istio · Linkerd · Consul. Cancel anytime.
The Problem
Service meshes are powerful but hard to automate
Service meshes like Istio and Linkerd give you powerful primitives for traffic management, observability, and security. But operating them requires deep expertise and every change means manual kubectl commands, YAML patches, or navigating complex dashboards. AI agents can't touch your mesh at all.
60%
Of Kubernetes engineers report service mesh configuration as their top operational complexity
45 min
Average time to diagnose a service-to-service connectivity issue without a unified interface
3×
Faster incident resolution when AI agents can inspect and adjust service mesh config in real time
The Solution
10 MCP tools for AI-driven mesh orchestration
TRUSS Service Mesh Bridge gives AI agents a unified programmatic interface for your entire service mesh. Discover services, manage traffic, monitor health, and orchestrate cross-service workflows — all via MCP tool calls.
Traffic Routing
Set traffic weights for canary deployments, A/B tests, or blue-green switches. Route by headers, cookies, or source service. No YAML editing required.
Service Metrics
Retrieve request rate, error rate, and latency percentiles (p50/p95/p99) for any service or endpoint. Get the data agents need to make traffic decisions.
Health Monitoring
Check the health status of any service, pod, or endpoint. Surface unhealthy instances, failed probes, and degraded replicas before they cascade into outages.
Circuit Breaker Control
Open, close, or half-open circuit breakers for any service. Configure thresholds for error rate, consecutive failures, and request volume. Prevent cascade failures programmatically.
Retry & Timeout Policy
Configure per-service retry budgets, timeout durations, and retry conditions. Set different policies for different route subsets. Tune resiliency without redeploying services.
Fault Injection
Inject delays or HTTP faults into specific service routes for chaos engineering and resilience testing. Verify that your circuit breakers and retry policies actually work.
Access Log Streaming
Retrieve and filter service-to-service access logs from Envoy or your mesh proxy. Filter by source service, destination, status code, latency threshold, or time window.
Multi-Service Orchestration
Execute a sequence of calls across multiple services in your mesh as a single logical workflow. Chain responses, handle errors, and maintain distributed transaction context.
Dependency Mapping
Generate a real-time dependency graph for any service showing all upstream and downstream dependencies, their current health, and the critical path for your request flow.
How It Works
From install to mesh control in 60 seconds
TRUSS Service Mesh Bridge is an MCP server. Install via npm, configure your cluster access, and AI agents can start managing your mesh immediately.
-
01
Install and configure cluster access
Install via npx. Provide your kubeconfig or mesh control plane URL. Supports Istio, Linkerd, and Consul Connect. In-cluster deployment available for production use.
-
02
Add to your MCP client
Add to Claude Code, Cursor, or any MCP-compatible agent. All 10 mesh tools appear immediately. No UI required.
-
03
Let agents orchestrate your services
Ask your agent to "shift 10% of traffic to the new cart-service version" or "diagnose why checkout-service is showing elevated latency." The agent calls the MCP tools and takes action.
# Claude Code / Cursor MCP config
{
"mcpServers": {
"mesh-bridge": {
"command": "npx",
"args": ["truss-mesh-bridge"],
"env": {
"MESH_TYPE": "istio",
"KUBE_CONFIG_PATH": "~/.kube/config",
"MESH_NAMESPACE": "production",
"TRUSS_API_KEY": "your-api-key"
}
}
}
}
Pricing
Simple, flat pricing
One plan. Unlimited service mesh operations. $49/mo flat.
Pro
$49/mo
Full service mesh control for one cluster environment.
- All 10 MCP tools
- Istio, Linkerd, Consul Connect
- Unlimited service operations
- Traffic routing & canary deploys
- Circuit breaker management
- Fault injection & chaos testing
- Multi-service orchestration
- Works with Claude Code & Cursor
Tool Reference
All 10 service mesh tools
- 01 discover_services — List all mesh services with versions, endpoints, health, and traffic weights
- 02 route_traffic — Set traffic weights for canary, A/B, or blue-green by header or service
- 03 get_service_metrics — Request rate, error rate, and latency percentiles for any service
- 04 check_health — Surface unhealthy instances, failed probes, and degraded replicas
- 05 manage_circuit_breaker — Open, close, or configure circuit breakers programmatically
- 06 set_retry_policy — Configure per-service retry budgets, timeouts, and conditions
- 07 inject_fault — Inject delays or HTTP faults for chaos engineering and testing
- 08 get_access_logs — Retrieve and filter Envoy access logs by service, status, or latency
- 09 orchestrate_workflow — Chain calls across services in a single agent-managed workflow
- 10 get_dependency_map — Real-time dependency graph with health and critical path analysis
FAQ
Frequently asked questions
What service meshes does the MCP bridge support?
TRUSS Service Mesh Bridge supports Istio, Linkerd, and Consul Connect. It uses standard Kubernetes and mesh control plane APIs, so it works with any Kubernetes-based deployment. Set MESH_TYPE in your config to select your mesh.
How does AI microservices orchestration work?
Your AI agent calls MCP tools to discover services, check health, adjust traffic, and chain multi-service calls. The agent reasons about your service topology and can handle complex workflows — like "shift traffic away from unhealthy instances, check if error rate drops, then alert if it doesn't" — that would otherwise require custom scripts.
Does it require changes to my existing services?
No. The MCP bridge connects to your existing service mesh control plane via standard APIs. Your services don't need modification. You need cluster access from the environment where the MCP server runs — either via kubeconfig or in-cluster service account.
Does it work with Claude Code and Cursor?
Yes. TRUSS Service Mesh Bridge is a standard MCP server. Add it to your mcpServers config in Claude Code, Cursor, or any MCP-compatible client. Your agents can start managing your service mesh immediately.
Give your AI agents control of your service mesh
Stop writing kubectl commands and YAML patches by hand. Let AI agents manage your microservices through a unified MCP interface.
Get Mesh Bridge MCP — $49/moFlat $49/mo. Istio, Linkerd, Consul. Cancel anytime.