Know how your team actually works with coding agents.
Flowlines shows adoption across instrumented Claude Code and Codex sessions: active days, task context, delegation, verification, retries, outcomes, and the developers behind each pattern.
01
The adoption question
Seat counts and licenses do not show whether developers actually use coding agents, return to them, delegate meaningful work, or verify what gets produced.
Flowlines measures observed sessions and active days by developer, then keeps the task context attached.
02
The work pattern
See cadence, reuse, automation, delegation, verification commands, retry or repetition flags, tool activity, and task delivery outcomes.
Compare developers or cohorts without turning one productivity metric into a score of the person.
03
The quality question
Open the sessions behind missing verification, repeated attempts, abandoned tasks, weak results, tool failures, or a claimed completion that lacks evidence.
Release and agent context show whether the pattern changed after tooling or policy updates.
04
How you connect it
For supported coding-agent CLIs, install the Flowlines plugin globally. It observes Claude Code and Codex sessions without changing application code. MCP servers use a separate OpenTelemetry instrumentation path.
look at the work, not just the seats ↙
Coding-agent adoptionInstrumented Claude Code and Codex activity