Spanly
See what AI agents do inside your MCP server
Last verified:
What is Spanly?
Spanly is a dedicated observability and monitoring platform for Model Context Protocol (MCP) servers. It provides real-time tracing, error tracking, and complete visibility into every interaction between AI agents and MCP servers, filling the gap where traditional APM tools only see HTTP traffic but not MCP-specific traffic.
Key features include real-time tracing of MCP requests, error tracking with alerting when error rates spike or latency degrades, team dashboards with custom views for different teams (DevOps monitors uptime, developers debug traces), audit logs for compliance and debugging, and near-zero SDK overhead of less than 0 ms per request. The platform scales from one MCP server to hundreds with high-throughput ingestion and flexible retention periods.
Spanly is designed for developers building AI agents, DevOps teams managing MCP infrastructure, and organizations running AI agent deployments that need to monitor MCP server traffic. It requires only 5 lines of code to start monitoring—install the @spanly/spanly npm package, wrap your MCP server with the SpanlyClient, and get full visibility instantly with no complex setup or config files.
The platform offers alerts for error rate spikes and latency degradation, team dashboards with custom views, audit logs for compliance, and is building future features like full-text search across traces/errors/sessions, data export for custom analysis, and REST API access for programmatic access to observability data.
Spanly pricing
Pricing model: Freemium
Free: $0/mo - 10,000 MCP requests/month, 14 days data retention, 1 seat, basic analytics, no credit card required. Starter: $29/mo - 200,000 MCP requests/month, 90 days data retention, 5 seats, advanced analytics, $4 per 100k additional requests. Business: $99/mo (Best Value) - 2,000,000 MCP requests/month, 1 year data retention, unlimited seats, priority support, $3 per 100k additional requests. Enterprise: Custom plans available with higher limits, dedicated support, custom retention periods, and SLA guarantees by contacting Spanly.
Spanly pros
- Dedicated MCP traffic observability layer alongside Datadog, Sentry, or New Relic
- Real-time tracing of every MCP request between AI agents and servers
- Error tracking with automatic alerting on error rate spikes
- Alerting on latency degradation beyond custom thresholds
- Team dashboards with custom views for DevOps and developers
- Audit logs for complete compliance trail of MCP interactions
- Less than 0 ms SDK overhead per request
- 0% uptime SLA for data ingestion
- Zero lines of additional code required beyond 5-line SDK wrap
- No complex setup or config files needed
- Scales from 1 to hundreds of MCP servers seamlessly
- High-throughput ingestion for large AI infrastructure
- Flexible data retention periods by plan (14 days to 1 year)
- Free tier with 10,000 requests/month and no credit card required
- Simple npm install with zero dependency headaches
- Prorated billing for upgrades, immediate effect
- Unlimited seats on Business plan
- Priority support included on Business plan
- Overage charged at $3 per 100k on Business plan
- Enterprise plans with custom limits and dedicated support available
Spanly cons
- Full-text search feature not yet available (coming soon)
- Data export feature not yet available (coming soon)
- API access not yet available (coming soon)
- Only 10,000 requests/month on Free tier
- Free tier only has 14 days data retention
- Free tier limited to 1 seat only
- Free tier has only basic analytics, no advanced analytics
- Starter plan only has 5 seats, not unlimited
- Starter overage is $4 per 100k, higher than Business
- No SLA guarantees mentioned for paid plans
- Monitoring pauses on Free plan when over limit until next cycle
- Downgrades only take effect at end of billing cycle
- Only npm package available, no other SDKs mentioned
- Requires API key setup environment variable
- Enterprise plans require contacting sales, no self-signup
Frequently asked questions about Spanly
What is included in the Free plan?
The Free plan includes 10,000 MCP requests per month with 14 days of data retention, 1 seat, and basic analytics. No credit card is required to start.
What counts as one MCP request?
Each MCP protocol message — such as an initialization, tool call, prompt execution, or resource access — counts as one request. The request and the response are counted as one MCP request together.
Can I upgrade or downgrade my plan at any time?
Yes, you can upgrade or downgrade your plan at any time. Upgrades take effect immediately and billing is prorated. Downgrades take effect at the end of the current billing cycle.
What happens if I exceed my request limit?
On paid plans, you are charged for overage at your plan's per-100k rate ($4 on Starter, $3 on Business). On the Free plan, monitoring pauses until the next billing cycle when you exceed the limit.
How long is data retained on each plan?
Data retention depends on your plan: 14 days on Free, 90 days on Starter, and 1 year on Business. Enterprise plans offer custom retention periods tailored to your organization.
Are enterprise plans available with custom limits?
Yes! Contact Spanly for custom enterprise plans with higher limits, dedicated support, custom data retention periods, and SLA guarantees tailored to your organization's needs.
How many lines of code do I need to start monitoring?
You need only 5 lines of code to start monitoring. Install the @spanly/spanly npm package, create a SpanlyClient with your API key, and call spanly.monitor(mcpServer) — no complex setup or config files required.
What SDK overhead does Spanly add per request?
Spanly adds less than 0 ms of SDK overhead per request, meaning virtually no performance impact on your MCP server operations.
Does Spanly work alongside my existing APM tools?
Yes, Spanly is the dedicated layer for Model Context Protocol (MCP) traffic that works alongside the Datadog, Sentry, or New Relic you already run. Your APM sees HTTP but Spanly sees MCP specifically.
How do alerts work in Spanly?
You get notified when error rates spike or latency degrades beyond your custom thresholds. Set up alerts to monitor your MCP server health and get real-time visibility into performance issues.