Guides

Using the dashboard

The dashboard is the visual interface to the same engine the CLI drives. It's where you watch runs move in real time, explore the memory graph, read analytics, and manage connectors, schedules, and your team — things a terminal can't show well.

Everything here runs the identical workflows described in Workflows — starting a run from a dashboard button and starting it from moa wf … hit the same server. Reach it at app.fnmoa.com and sign in with GitHub.

Your first five minutes#

A quick orientation loop: see work, drill in, then trust the numbers.

Open the Overview to see active and recent runs at a glance.
Click a run to open its timeline and watch the steps progress.
Check Intelligence for the rates and hotspots across all your runs.

Overview — the activity view#

The landing screen. It shows your workflow runs and where each one is — queued, running, or finished — so you can tell what Moa is doing across your repos right now. A stage moving forward means the deterministic harness advanced the run; a stop means the run finished, failed, or was canceled.

Run timeline#

Open any run for its live step-by-step timeline — branch, agent, verify-gate, open-PR, and so on — with timing and the outcome of each step. This is the canonical record of a run: what the agent reasoned about and every git action the harness took on its behalf. Failed runs can be resumed (replaying cached steps so the agent doesn't redo finished work) or restarted from scratch.

Dispatch — start a run#

Kick off any workflow by hand: pick a repo and an issue or PR, choose the workflow, and go. Useful when you want to solve, review, or fix-CI something without waiting for a label or webhook to trigger it.

Memory graph & search#

A visual, explorable graph of everything Moa has learned, powered by LocusGraph. Expand a node to see how a run, a file, and a past failure connect, and use natural-language search to ask the graph questions in plain English. This is the memory behind recall-before-fix — and it's isolated per org.

Intelligence — analytics#

Aggregate signal across all your runs. Read each number as what it is → why it matters:

CI-fix rate — how often fix-ci turned a red build green. A falling rate is an early sign of a deeper or flaky failure worth a human look.
File hotspots — which files change most across runs. A high-churn file is a candidate for refactoring or extra review.
Trends — run volume and outcomes over time, so you can see whether Moa is getting more or less effective on your repos.