Screenshots

The unfirehose dashboard running locally. Every session, every thought thread, every token — captured and visualized.

Real-time streaming of agent sessions. Watch thought tokens, tool calls, and diffs as they happen across all harnesses.

Grid view of all running coding agent sessions. Model, message count, token usage, and duration at a glance.

Activity overview with session counts, token totals, cost tracking, hourly distribution, and day-of-week heatmaps.

Like Last.fm for code. Activity heatmap, hour-of-day distribution, daily cost tracking, streak counter, and lifetime stats.

Browse extended reasoning traces from machine learning models. Search across all thinking blocks. The full chain of thought preserved.

Breakdown by model, provider, and type. Input vs output, cache read vs write. Per-model cost tracking with efficiency ratios.

All your repositories with session counts, last activity, and git status. Navigate to any project's session history.

Per-project session timeline with file tree, git history, and message drill-down.

Cross-session todos extracted from all agent conversations. Kanban board with pending, in-progress, and completed lanes.

Percentile latency tracking across models. Compare response times, monitor degradation, spot provider issues.

Real-time token consumption with configurable thresholds. Alerts when spending exceeds your budget window.

Mesh node overview. CPU, memory, disk, GPU, power consumption across your compute fleet. Economics dashboard included.

Permacomputer mesh with node health, running harnesses, and tmux session monitoring across all machines.

The unfirehose/1.0 specification. Object types, harness adapters, field mappings. The standard that connects every coding agent.

Dashboard configuration, API key management, cloud sync settings, and plan selection.

See It Running

Free. Open source. AGPL forever. Self-host in 60 seconds.