See the use case
Understand the real jobs people and agents give the server.
MCP observability
Know what people and agents do with your MCP servers.
Flowlines shows server inventory, tool adoption, clients, reported identities, complete use-case journeys, outcomes, loops, failures, and the exact calls and sessions behind them.
Servers and tools
Start with the full MCP estate.
See which servers are active, which tools drive adoption, who calls them, where call reliability changes, and whether outcome context is complete.
MCP serversObserved server inventory
OpenTelemetry tools/call spans · last 7 daysSelected server
Receiving callscustomer-context-mcp
Calls24,814Sessions844Reported identities286Outcome coverage93%Failed calls1.2%
search_accounts8,21414299.3%
enrich_company5,10811998.8%
get_signals4,8129797.9%
write_fields2,2646194.2%
What call status misses
write_fields succeeds at the protocol layer, but 11 journeys leave requested fields unchanged.
Journey reconstruction
A successful tool call is not a successful journey.
Reconstruct the use case from request to outcome, including the client, reported identity, tool path, repeated calls, missing results, and the sessions affected.
MCP journeysUse case to outcome
Example workspace · last 7 daysThemed sessions748Reported users286Use-case journeys12Loops to review30
USE CASEsearch_accountsTOOLenrich_companyTOOLget_signalsOUTCOMEBrief ready
7 repeated-call loops
Eight journeys ended without a reported outcome. Seven repeated the same signal lookup.
From alert to cause
Follow the failure back to the session.
Follow the journey
See the typical tool path, loops, failures, and reported outcome.
Find the people
Know which users, agents, and clients drive each behavior.
Open the evidence
Inspect the original request and tool calls behind the pattern.
Before you connect
MCP setup, explained.
What does MCP observability show?
Flowlines shows observed MCP servers, tool adoption, calls, reported identities, clients, use-case journeys, repeated-call loops, failures, outcome coverage, and evidence sessions.
Is MCP latency the main Flowlines metric?
No. Latency and call status remain useful, but Flowlines focuses on what people and agents do with the server, whether the complete journey worked, and who was affected.
Does MCP observability require instrumentation?
Yes. The MCP server must emit a canonical tools/call span. The Flowlines skill can instrument compatible repositories through AGNTCY Observe or standard OpenTelemetry after the export boundary is reviewed.
Can Flowlines identify MCP users?
Flowlines can group calls by a person, agent, workspace, or MCP client when the server reports a stable permitted identity.
MCP server setup
Add MCP observability from Claude Code or Codex.
The Flowlines skill reviews the server, asks what may be exported, then adds canonical tool-call telemetry with AGNTCY Observe or OpenTelemetry.
Choose your coding agent
Open the setup where you already work.
One click opens a guided instrumentation task. You approve every change.
The skill finds the framework, tool handlers, and existing telemetry.
You approve which arguments, results, and identities may leave the server.
The skill makes the smallest compatible OpenTelemetry integration.
Run one tool flow and confirm its intent, calls, and outcome in Flowlines.
Prefer the terminal?Show one-line commands
Claude Code
claude plugin marketplace add flowlines-ai/plugins && claude plugin install flowlines@flowlinesCodex
codex plugin marketplace add flowlines-ai/plugins && codex plugin add flowlines@flowlinesThen use /flowlines:flowlines-mcp-observability in Claude Code or $flowlines-mcp-observability in Codex. Source and release history are in the official Flowlines plugin repository ↗.
From unknown to visible