Runs the operations of your Web3 Discord: rewards and payouts, ambassadors and KOLs, regional teams, content and reports. Your team approves everything that touches money, roles and people.
Core built and battle-tested on a live multi-regional Web3 program.
You start with one workflow, the one that bleeds most. Pick your machine.
Ombra fits some teams and not others.
FITS
A paid ambassador, creator or KOL program with a recurring reward cycle
Regional leads or moderators you have to chase for updates
Five or more people doing community operations
One person who owns the process and can approve decisions
DOES NOT FIT
✕You need moderation and spam cleanup only
✕You want a bot for one campaign, then done
✕Nobody owns the process on your side
✕Your community operations are not recurring yet
If you are in the right column, we will say so on the call instead of selling you a pilot.
PACKAGE A · FOR PROJECTS WITH A PAID AMBASSADOR OR CREATOR PROGRAM
Payout day becomes an hour of review.
Collects contributor content from your channels and tracker on your cycle schedule
Flags bots, farms and low-effort AI content, with evidence. You decide, nothing is auto-punished
Drafts the payout sheet by your criteria, marks spikes, duplicates and farm patterns
You hit approve. It publishes results and DMs every contributor their status
Every "why wasn't I paid" closes with an audit-log receipt
KOL vetting before you pay the invoice: how many of those "200 referrals" are real people
{{ pct }}%
of accounts that had already passed manual review turned out to be junk.
One live KOL campaign, 200 submitted referrals, junk = fresh wallets under 7 days or wallet-cluster matches. One program, shown as an example, not a promised detection rate.
A weekly summary for you, a monthly report for leadership. Every number is produced by the system itself and carries an audit line: which data, which rule, who confirmed.
"What am I paying for?" becomes a one-minute answer.
Week 12 summary
18/20 check-ins on timeaudit ↗
2 flags reviewed, 1 payout correctedaudit ↗
1 region alert, reaction in 4haudit ↗
payout cycle: 1h 10m reviewaudit ↗
Every number carries an audit line
Won't an AI wreck my server right before TGE?
No. Here's how it's built.
It starts read-only: it touches nothing while it maps your cycle, and hands you the audit in week one
Money, roles and people: a human approves, always
Every action logged, silent edits surface publicly
Isolated deploy: your server, your database, your keys
We never read private DMs and never train on your community's messages.
Founding Pilot
LIMITED SEATS
$1,000 setup + $750/mo· cancel any time
Week 1 you get the audit. Week 2 it runs live on your approvals. After that it is monthly, and the numbers decide.
Included: Week-one audit: a map of where your cycle leaks, plus a retroactive farm and junk scan of your last payout. One package, one data source, one approver. Weekly and monthly reports. Criteria changes included. Isolated deploy.
Later: working clients can add the second package, snapshot role-sync, extra regions and languages.
A junior assistant costs the same, learns for two months, makes mistakes in sheets where errors cost tokens, and takes the process with them when they leave. Ombra leaves your project the process, the history and the audit log.
A dashboard shows a chart when you open it. Ombra comes to you with the drop, the silent lead and the next step, and chases the case until it has an owner.
We have Zealy / Galxe.
Quest platforms generate submissions and basic anti-sybil. Nobody reviews quality or runs the payout cycle end to end. Ombra works on top of your tracker, Zealy included.
Our criteria change every month.
Criteria live in a config, changes are included in the subscription. That's why this is a managed service, not a box.
How accurate is the farm detection?
Scoring is probabilistic. Ombra gives you a shortlist with evidence attached, and you make the call. We do not publish an accuracy rate until we have run on your data, because a number from someone else's community would tell you nothing about yours.
How fast do we see results?
Check-ins, flagging and payout drafting work from the first cycle. Trend lines need two to four weeks of history before they mean anything. That distinction is written into the contract.
An AI with server access sounds risky.
It starts read-only. Permissions are scoped to the package. Everything risky waits for your yes. Every action is logged.
Does Ombra replace our community manager?
No. It removes the recurring manual work from their week: collecting, filtering, drafting, chasing, reporting. Decisions about money, roles and people stay with your team, and the page above shows exactly where the approval step sits.
What access does it need in Discord?
It starts read-only. After that, permissions are scoped to the package you picked, nothing wider. You can revoke access at any time from your own server settings, and every action it ever took stays in the audit log.
How long does setup take, and what do we have to do?
One week from signature to the audit landing in your inbox. From your side: one kickoff call, access to the channels and the tracker, and your reward criteria in whatever form they exist today. We turn them into the config.
Can we stop after the first month?
Yes. There is no minimum term and no notice period. The setup fee covers the week-one audit, which is yours to keep either way, along with the config, the process documentation and the full audit log.
Who owns the data and the process?
You do. Isolated deploy on your side, your database, your keys. We never read private DMs and never train on your community's messages.
Show us your workflow. We'll tell you where it breaks.
30 minutes: we walk your reward cycle or your team week, find the failure points, and say honestly whether we'd take it on.
You leave with a map of where your process leaks, whether or not we work together.