tailflow
v0.3.3
Published
Runtime verification for coding agents using local process, Docker and file logs over MCP
Maintainers
Readme
TailFlow
Runtime verification for coding agents.
TailFlow turns output from local development processes, Docker containers, and log files into bounded evidence an agent can use to verify a code change.
It helps an agent:
- confirm that expected services actually started;
- collapse repeated failures while retaining stack-trace context;
- wait for rebuilds and other asynchronous outcomes without polling; and
- compare new runtime output against a cursor captured before its edit.
npm install -g tailflow
tailflow init
tailflow-daemon
claude mcp add tailflow -- tailflow-mcpThe package installs four interfaces over those capabilities:
tailflow-daemoncollects logs and serves the agent API and dashboard;tailflow-mcpconnects an MCP client to the daemon;tailflow-logsqueries the daemon from a shell or script; andtailflowopens the interactive terminal UI.
Read the quick start and project overview or browse the feature guide and complete documentation.
TailFlow is a local runtime-verification layer. It is not a hosted log platform or a replacement for production observability.
