Docs

Orchestrator

The Orchestrator is Fload's AI brain — it coordinates all other agents, generates daily briefings, and decides what actions to take based on what's happening across your apps.

Think of it as an operations manager that never sleeps. It observes your metrics, reviews agent outputs, makes decisions, and dispatches work — multiple times a day.


How it works

The Orchestrator runs on a configurable schedule (default: every 6 hours) and follows a four-phase cycle:

1. Observe

Collects results from all recent agent runs (review replies, growth audits, ad performance, anomalies). Pulls the latest metrics from every connected data source.

2. Review

Analyzes everything together. Identifies patterns and issues that span multiple areas — for example, a spike in negative reviews combined with a drop in retention might signal a bug in your latest release.

3. Decide

Prioritizes what needs attention. Considers what worked (and what didn't) from past actions using its built-in memory system. Ranks actions by impact and urgency.

4. Dispatch

Creates pending actions for your approval — or auto-executes them if you've configured it that way. Queues specialized agent runs as needed (e.g., "run a growth audit" or "generate new ASO recommendations"). Produces the daily briefing.


Daily briefing

Every cycle produces a briefing that includes:

  • Key metrics summary — what changed since the last briefing
  • Agent activity — what each agent did and found
  • Anomalies & alerts — anything unusual that needs attention
  • Recommended actions — what the Orchestrator thinks you should do next
  • Pending actions — actions waiting for your approval

Briefings are delivered via your preferred channel — email, Slack, or Discord.


Memory and learning

The Orchestrator remembers past decisions and their outcomes:

  • After dispatching an action, it schedules a check-back to evaluate results
  • Successful actions inform future decisions ("pausing that campaign last time improved ROAS")
  • Unsuccessful actions are flagged so the same mistake isn't repeated
  • Over time, the Orchestrator gets better at understanding your specific apps

Event-triggered mini-cycles

Beyond scheduled runs, the Orchestrator reacts to urgent events:

  • Anomaly detected — a metric spiked or dropped significantly
  • Negative review surge — multiple 1-star reviews in a short period
  • Ad spend anomaly — unexpected budget burn or campaign failure
  • Submission rejection — your app was rejected by App Review

Mini-cycles follow the same observe-review-decide-dispatch flow but focus on the triggering event.


Draft vs Agent mode

The Orchestrator operates in two modes:

Draft mode

The default. The Orchestrator generates recommendations and queues them as pending actions for your review. Nothing changes until you approve it. This is ideal when you're getting started or managing high-stakes apps.

Agent mode

The Orchestrator acts on its own — executing decisions without waiting for approval. It still logs everything and respects safety guardrails (like budget caps), but it moves faster. Best for teams that trust the system and want hands-off operations.

You can switch between modes at any time in Settings → Orchestrator.


Agent write actions

When the Orchestrator decides something needs to happen, it coordinates write actions across agents:

  • Review Agent — post a reply to a review
  • ASO Agent — update promotional text or descriptions
  • Ads Agent — pause a campaign, adjust a bid, or reallocate budget

All write actions are tracked in a unified activity log. In draft mode, they appear as pending actions. In agent mode, they execute and you're notified after the fact.


What it can do

The Orchestrator's available actions depend on your connected platforms:

Connected platformWhat the Orchestrator can coordinate
App Store ConnectReview replies, ASO updates, submission checks
Google PlayReview replies, description updates
Apple Search AdsPause/resume campaigns, adjust bids and budgets
Google AdsPause/resume campaigns, adjust bids and budgets
Meta AdsPause/resume campaigns, adjust bids and budgets
TikTok AdsPause/resume campaigns, adjust bids and budgets
Stripe / RevenueCatRevenue monitoring, tier alerts
Slack / DiscordNotifications, briefing delivery

The more platforms you connect, the more the Orchestrator can do.


Configuration

Run frequency

  • Default: every 6 hours (4 times daily)
  • The schedule is configurable
  • Mini-cycles run on-demand regardless of schedule

Approval mode

  • Draft mode — all actions require your manual approval before executing
  • Agent mode — all actions execute automatically (requires explicit opt-in)

Notifications

  • Choose which events trigger notifications
  • Pick your delivery channel (email, Slack, Discord)

What's free vs. paid

FeatureFreePro ($50/month)
View daily briefings and agent activityYesYes
View pending actions queueYesYes
View activity log and historyYesYes
AI-powered Orchestrator cyclesYes
Automated agent dispatchYes
Event-triggered mini-cyclesYes

See Pricing for the full breakdown.

PreviousHow Fload Works
NextReview Agent