~60 min/day
ROAS checks & bleeder pauses
Open every campaign, eyeball the curves, pause what's bleeding — and hope you caught it before lunch, not after.
Your Shopify storeruns on
// powering the next wave of dtc
Brand names shown are illustrative composites.
Reef & RangeNORTHWOODLinen HouseSOLACEHatley + CoMODULARStoke GoodsBREVARDEmber & Co.OKAY WILDDrumrollPLAINFIELDReef & RangeNORTHWOODLinen HouseSOLACEHatley + CoMODULARStoke GoodsBREVARDEmber & Co.OKAY WILDDrumrollPLAINFIELD
// the daily grind
Running a store well costs 4–6 hours a day. Every day.
Not strategy — upkeep. The recurring checks and fixes that keep revenue from leaking. Magistry is built to take exactly this work off your plate.
~30 min/day
Merchant Center & feed fires
Disapprovals, policy flags, stale feeds across five channels. Each one quietly throttles your reach until someone notices.
~90 min/day
The CS inbox
Where's-my-order, sizing complaints, refund requests. Every reply needs the order, the policy, and your brand voice.
~60 min/day
Catalog & pricing housekeeping
Draft the dead stock, reprice the movers, vault the seasonal items, watch for products cannibalizing each other.
Agents run this loop continuously — overnight, on weekends, across every store you own. You review decisions, not dashboards.
// the platform
One platform. Every job your store runs on.
Not point tools — one operating system across growth, operations, and defense, with every external write on the same auditable spine.
// grow revenue
Ads, feeds, and tracking that act on profit — not vibes.
The Campaign Agent paces budgets and bids on Google Ads, briefs and generates creative, and writes Customer Match audiences. Your shopping feed syndicates to Google, Meta, TikTok, Pinterest, and Bing with rules, versioning, and A/B arms. Underneath it all: server-side tracking with an identity graph, multi-platform conversion delivery, and POAS bidding that optimizes on profit, not just revenue.
- Budget & bid pacing with anomaly auto-pause
- Shopping-feed syndication to 5 channels + Merchant Center auto-fix
- Server-side tracking, attribution & POAS bidding
- RSA copy + ad-image generation, pushed on approval
channels · live writes
Google Ads
ROAS 3.8
Shopping feed ×5
synced 12m
Server-side CAPI
4 platforms
POAS bidding
profit-true
campaign_agentADJUST_BUDGET · EU/Linen/PMax · +12% headroom
// run lean
A catalog and inbox that manage themselves.
The Catalog Agent runs every SKU's lifecycle — drafts underperformers, publishes winners, prices with margin gates, vaults seasonal items and revives them six weeks before peak, and catches products cannibalizing each other. The CS Agent answers inbound end-to-end: policy-grounded drafts in your brand voice, refunds and replacements gated by expected value, escalation when it's not sure.
- Lifecycle automation with self-calibrating thresholds
- Margin-aware pricing — only on verified cost data
- CS replies grounded in your policies, scored against your voice
- Refund / replace / gift-card execution with EV gates
lifecycle · last 24h
- DRAFTSKU-9241 · ROAS < floor↶
- PUBLISHSKU-1182 · new-arrival scout↶
- SEASONAL_VAULT12 SKUs · revive wk 38↶
- ISSUE_REFUNDorder #38502 · EV-positive↶
margin gates · tier Aself-calibrated thresholds
// protect & discover
Defense and discovery, running in the background.
Dispute Shield assembles chargeback evidence and submits it. Reputation monitoring ingests reviews and mentions, classifies them, and drafts responses — with a crisis command center when things spike. The Market Expert watches news, social, and search demand for credibility-scored signals, while the Researcher scans seven lanes for your next winning product.
- Chargeback evidence assembly & submission
- Review / mention monitoring with drafted responses
- Market signals from news, social & search trends
- 7-lane product research with supplier matching
defense & discovery
Dispute Shield
evidence pack · 9 exhibits
submittedReputation
3 mentions · 1 review
draft repliesmarket_expertlinen demand spike · EU · credibility 0.81
researcher14 candidates · 7 lanes · top score 86/100
// how it works
From install to autonomy — at your pace.
Autonomy is earned, not assumed. Every agent starts read-only and only widens as far as you let it.
Connect Shopify
Install from the Shopify admin in minutes. Magistry syncs your catalog, orders, costs, and ad accounts — secrets vault-encrypted from the first byte.
OAuth install · no code · no credit card
Watch the dry-run
For the first 24 hours every agent runs in dry-run: decisions land in the log with full reasoning and evidence, but nothing is written to your store.
decision_log fills · zero writes
Approve, then go autonomous
Start in approval mode and accept decisions one by one. When the rows keep matching your gut, widen autonomy per agent — and roll back any row, any time.
per-agent modes · ↶ rollback · kill switch
Start free
Free plan · No credit card · First 24 hours run in dry-run
// the fleet
Four specialists — ready to work.
Not a chat wrapper. Not a workflow builder. Each Magistry specialist is a planner-judge-executor loop with eight specialist skills wired into your real systems — and they share signals with each other.
Catalog Specialist
// Runs the lifecycle of every SKU.
Drafts underperformers, archives dead stock, runs margin-aware discounts, and publishes winners — every product in exactly one lifecycle state with a reversible row attached.
// writes to
Shopify Admin GraphQL
// actions
DRAFTARCHIVEPRICE_CHANGECLEARANCEPUBLISH
Campaign Specialist
// Operates your paid acquisition.
Eight specialist skills — bid pacing, audience writer, creative director, anomaly responder, compliance — wired into a planner-judge-executor loop per campaign and channel.
// writes to
Google Ads Meta TikTok Bing
// actions
ADJUST_BUDGETBID_STRATEGYPAUSE / RESUMENEGATIVE_TERMSAUDIENCE_WRITECREATIVE_BRIEF
CS Specialist
// Handles inbound end-to-end.
Order Detective + Policy RAG + Reply Drafter + Reply Judge. Brand-voice embeddings hold tone; language-gated send; proactive outbound on in-flight issues before the customer asks.
// writes to
IMAP / SMTP Shopify Orders Refund APIs
// actions
ISSUE_REFUNDISSUE_REPLACEMENTCHANGE_ADDRESSISSUE_GIFT_CARDREPLY_DRAFTESCALATE
Market Expert
// Watches the market for you.
Scans news, social trends, and search demand across research lanes; publishes credibility-scored signals the other agents subscribe to — demand spikes, niche refreshes, geo-expansion windows.
// writes to
News / GDELT Social trends Keyword intel
// actions
MARKET_SIGNALNICHE_REFRESHGEO_EXPANDCREDIBILITY_SCORE
Agentsthe fleet
⏻ kill switchCatalog Agentautonomous
PRICE_CHANGE · linen blazer · +$4
31 decisions · 24h
Campaign Agentautonomous
ADJUST_BUDGET · EU/Linen/PMax · +12%
9 decisions · 24h
CS Agentapproval_only
REPLY_DRAFT · order #38241 · refund EV +$31
17 drafts · 3 pending
Market Expertdry_run
SIGNAL · linen demand spike · EU · cred 0.81
4 signals · 24h
cross-agent signals · 24h
- market_expert→campaigndemand_spike (EU) → budget headroom check
- cs→catalogcomplaint_cluster (sizing) → variant review
One gallery, four agents — each with its own autonomy mode, scorecard, and kill switch.
// audit plane
Audit — every decision grounded in evidence you trust.
Multi-system writes, idempotent rollbacks, append-only logs, and continuously-tuned guardrails — unified in one auditable plane the operator owns.
Shift inventory state with daily action plans
Lifecycle transitions across the catalog, validated against ALLOWED_TRANSITIONS, applied in batches with reversal pre-computed.
Allocate ad spend with real-time decision evidence
Budget pacing and bid moves cited by anomaly windows, judged against policy, gated by tier-of-evidence before any euro moves.
Prevent runaway autonomy with kill switches
Default-on kill switch, per-action rate limits, advisory locks, vault-encrypted secrets, brand & trademark gates on generated copy.
Measure operator trust over time
Decision-row acceptance, dry-run-to-live ramp curves, judge agreement rates — surfaced as the metrics that tell you when to widen autonomy.
Prove every move with append-only logs
decision_log, ad_decision_log, cs_thread_evaluations — every transition is a row with reason, evidence, citation, and the reversal op pre-stored.
decision_log · row #84193
agent : campaign
action : ADJUST_BUDGET
target : campaign 'EU/Linen/PMax'
delta : +12% daily cap (€420 → €470)
reason : 14d window — CPA −18%, conv +9%
headroom verified by anomaly responder
policy mode = autonomous (above threshold)
evidence : [perf_window#9123, anomaly#771, judge_score=0.92]
applied_at : 2026-05-25T09:14:02Z
applied_by : campaign_agent.bid_pacing
reversal : ADJUST_BUDGET delta=-12% (one-click)
status : APPLIED
audit_chain : signal → plan → judge → execute → log↶ rollback availablestored forever (append-only)
// the product
Meet Otto.
Your autonomous operator, built specifically for ecommerce. It already knows your store, your margins, your brand voice — and now it does the work for you.
- Reads your config, your orders, and your decisions
- Runs a planner-judge-executor loop per agent
- Writes to live external systems with reversible rows
- Stays inside your kill switch, rate limits, and tier gates
Ottoin-app assistant
writes need approvalWhich products should I discount before the summer drop?
Three SKUs are sitting on Tier-A verified margin with falling sell-through: wide-leg trouser, knit polo — navy, belted shirt dress. A 14-day −15% clears projected dead stock before the drop while staying margin-positive.
PROPOSE_DISCOUNT · 3 SKUs · −15% · 14dpending
evidence: margin tier A · sell-through −34% · seasonal curve
ApproveEditreversible · logged
Otto reads everything, proposes writes — and writes only what you approve.
// difference
How is Otto different from a chatbot?
Most “AI for ecommerce” stops at suggestions and dashboards. Magistry is a teammate for everyday work — with the rails, evidence, and context to take action.
01
It does the work, not just the talking
Otto writes to Shopify, Google Ads, Meta, TikTok, and your inbox — directly. Not a chat wrapper. Not a CSV exporter. Every action lands in the system that runs your store.
PRICE_CHANGE · linen blazer · –€6 · tier-A · applied
02
It carries the full context of your store
Before every move, Otto reads your config, your decisions, your evidence, and your guardrails. It picks up exactly where last night's run left off — no warm-up, no re-prompting.
context: 84,193 decisions · 12 stores · 8h since last run
03
It turns any decision into an automation
Approve a decision row once, and Otto runs the same pattern — audited, gated, reversible — on a schedule. No glue code, no Zaps, no copy-paste from a chat transcript.
automation #142 · weekly · tied to decision_log #84102
// customers
Don't take our word for it.
Operators using Magistry to make every decision faster, evidence-backed, and reversible.
“AI is only as smart as the data it sees. Magistry has the full context of our store — orders, refund history, margin tiers — so its decisions actually map to ours.”
Maya Chen
Head of Ecommerce
Linen House
“If you're not running an autonomous control plane in 2026, you're paying three salaries to do what one row in a decision_log can do.”
Tomas Olsen
CEO
Reef & Range
“Magistry doesn't just recommend. It acts, in our accounts, inside our rules. That's the difference between a chatbot and a teammate.”
Jenna Park
Director of Growth
NORTHWOOD
“The decision_log is the killer feature. Every move is a row we can audit and reverse. My CFO sleeps better.”
Jacob Dorian
Head of Paid Media
SOLACE
“Our CS load dropped 40% the week we turned on the CS Agent. The remaining 60% is the work humans should actually be doing.”
Ben Diamond
COO
Stoke Goods
“We tried five workflow tools before Magistry. It's the only one we trust to write — and the rollback button is the reason.”
Sara Khan
VP Operations
Ember & Co.
“You don't scale a brand by stacking headcount. You scale it by giving the rails to a teammate that never sleeps and always logs.”
Victor Velazquez
Founder
MODULAR
“Magistry's Campaign Agent unlocked a 213% lift in new-customer acquisition in our first quarter. The numbers do the talking now.”
Alex Stark
CMO
OKAY WILD
“Otto answers my CFO's questions before they hit Slack. That's what 'AI for ecommerce' was supposed to mean from day one.”
Emma Liu
Head of Retention
BREVARD
Composite testimonials — illustrative of patterns we see across deployments, not specific named customers.
// case study · solace bands
4.3×
Faster decision throughput from manual approvals to autonomous execution.
Solace Bands moved 60+ external endpoints under one auditable spine. The result: less firefighting, less context-switching, more time on the bets that matter.
4.3x
Faster decision throughput
82%
Of decisions executed without operator touch
8m
Average time from anomaly to applied action
Composite example — day-one metrics, illustrative of patterns across deployments.
Read the full case studyHaving Catalog, Campaign, and CS agents writing in sync means our team finally focuses on strategy. Magistry runs the execution — every move is a row I can audit, reverse, or replay.
JD
Jacob Dorian
Head of Paid Media
SOLACE
// integrations
Connects in minutes. Writes with rails.
Per-store OAuth for every channel. Tokens live in an encrypted vault — never in plaintext — and every outbound write passes the same rate limits and kill switch.
- Shopifycatalog · orders · billing
- Google Adscampaigns · audiences
- Merchant Centerfeed · diagnostics
- Google Analytics 4sessions
- Metaads · CAPI · social
- Instagramposts · DMs
- TikTokfeed · CAPI
- Pinterestfeed · CAPI
- Microsoft Bingshopping feed
- Email (IMAP/SMTP)customer service
// more updates
Three more autonomous surfaces — wired in, ready to flip on.
NEW · v1.4
Social Autopilot
IG · FB · TikTok · X · LinkedIn · Pinterest. Post generation, image variation, DM responses, and an attention queue that ranks every reply by EV before it answers.
See Social AutopilotUPDATED · v1.3
Disputes & Reputation
Chargeback evidence packets assembled from the decision_log, fraud scoring at the order edge, and review-sentiment monitoring with auto-reply on signal — reversible when the dispute closes.
See dispute workflowsLIVE NOW
Researcher & Onboarder
Meta Ads Library, AliExpress trends, competitor bestsellers → frequency clusters → reverse-image supplier match → LLM copy → autopriced drafts ready for one-click publish.
“Researcher surfaces winning products before the trend hits Reels. We've onboarded 28 SKUs from zero to PDP in under 9 days.”
Marcus Lee · Founder · OKAY WILD
// plans
Every feature on every plan. Scale by usage, not paywalls.
All paid tiers unlock the whole platform — you grow by order volume and AI-credits. And there's no trial to expire: the Free plan is your trial.
Free
Free
50 AI-credits/mo · Up to 150 orders / mo
- Full dashboard (read-only)
- Every agent's recommendations — propose-only
- Decision log + audit trail
Starter
$39/ mo
500 AI-credits/mo · Up to 500 orders / mo
- Everything, fully unlocked
- Autonomous agents + write-actions
- Kill switch + per-action rate limits
Growth
$99/ mo
1,500 AI-credits/mo · Up to 2,500 orders / mo
- Everything in Starter
- 1,500 AI-credits / month
- Priority support
Scale
$249/ mo
4,000 AI-credits/mo · Up to 7,500 orders / mo
- Everything in Growth
- 4,000 AI-credits / month
- Priority support + dedicated CSM
// faq
Frequently asked questions.
Two categories — the product and the audit plane underneath it. Still unanswered? Book a 20-min call.
// magistry
About Magistry
// audit plane
About the Audit Plane
// magistry.io
Put your Shopify store on autopilot — with the rails on.
See what Magistry can do for your store. Connect Shopify in minutes. Run 24 hours in dry-run. Flip live when the decision rows match your gut.
Start on the Free plan · No credit card · No trial to expire · Kill switch ON by default