Multi-agent marketing orchestration

Agents that work for you.

10 specialized AI agents monitor your PPC accounts 24/7, propose concrete changes for approval and write monthly reports for you. 12 hours in Google Ads replaced by 90 minutes of review.

No credit card 5-minute setup Your data stays yours
workflow · monthly_audit RUNNING
run_1247 · step 1/9 · 12ms
Memory · 5 layers
Pipeline · 6 agents + Judge
Tools
L1 · Core
always in prompt
R/W
L2 · Episodes
approved/rejected
L3 · Patterns
Bayes · 146 concepts
R/W
L4 · Graph
Memgraph · multi-hop
L5 · Autodream
consolidation
1
Data Analyst
monthly_data_analyst
sonnet-4-6 · 16 384 tok
● running
2
Strategist
monthly_strategist
sonnet-4-6 · 16 384 tok
3
Structure Auditor
structure_auditor
sonnet-4-6 · 16 384 tok
4
Negative KW Agent
neg_kw_agent
sonnet-4-6 · 12 288 tok
5
Ad Copy Writer
ad_copy_writer
sonnet-4-6 · 16 384 tok
6
Report Writer
monthly_report_writer
sonnet-4-6 · 16 384 tok
Evaluator · Judge
missing point · refill · approve
Data tools 12 functions
get_budget_dataget_metrics_dataget_account_dataget_campaign_structureget_change_historyget_search_terms
Memory tools R/W 5 layers
core_memory_updatecore_memory_appendcore_memory_replacesearch_recallquery_memorygraph_query
Knowledge tools rules · benchmarks
get_knowledge_baseget_recent_insightscore_frameworkplatform_rulesclient_targetsanalysis_benchmarks
Action tools proposals · execute
draft_proposalexecute_changelog_actionnotify_clientget_current_goalget_success_prediction
AUTODREAM · micro
After each action
sorts & cleans episodes → L2 · L5
AUTODREAM · weekly
Monday full run
pattern, graph & core consolidation → L3 · L4 · L5
01 A1 · Data Analyst · reads L1+L3, calls Data+Memory tools

Connected to the systems where your budget actually moves

Google Ads
Sklik
Meta
LinkedIn
GA4
Search Console
Shoptet
Heureka
Client dashboard

Every client. One dashboard. Real time.

MTD spend, revenue, PNO, pacing and month-end forecast – per client, per platform. You see the story before your first coffee.

app.suryatool.io/dashboard
April 2026 · day 23/30 All clients
Total spend MTD
876 500
+12,4 %
Total revenue MTD
6,50M Kč
+18,1 %
Average PNO
13,5%
−1,8 %
Active clients
4
+1
Client / platform Spend MTD Plan Revenue MTD PNO Budget pacing Should be Forecast
CK Velký Orient 1 platform 268 500 Kč 340 000 Kč 1,94 M Kč 13,8 %
+3,0 %
260 667 Kč 350 217 Kč +341 Kč/den
Google Ads 1 účet 268 500 Kč 340 000 Kč 1,94 M Kč 13,8 %
+3,0 %
260 667 Kč 350 217 Kč +341 Kč/den
Farma Zelený Důl 3 platforms 413 400 Kč 560 000 Kč 3,35 M Kč 12,3 %
-3,7 %
429 333 Kč 539 217 Kč −693 Kč/den
Google Ads 1 účet 315 600 Kč 420 000 Kč 2,62 M Kč 12,0 %
-2,0 %
322 000 Kč 411 652 Kč −278 Kč/den
Sklik 1 účet 56 400 Kč 80 000 Kč 410 000 Kč 13,8 %
-8,0 %
61 333 Kč 73 565 Kč −214 Kč/den
Meta Ads 1 účet 41 400 Kč 60 000 Kč 320 000 Kč 12,9 %
-10,0 %
46 000 Kč 54 000 Kč −200 Kč/den
Statek Podlesí 1 platform 117 300 Kč 180 000 Kč 521 000 Kč 22,5 %
-15,0 %
138 000 Kč 153 000 Kč −900 Kč/den
Google Ads 1 účet 117 300 Kč 180 000 Kč 521 000 Kč 22,5 %
-15,0 %
138 000 Kč 153 000 Kč −900 Kč/den
Bylinná hořčice Opát 2 platforms 77 300 Kč 95 000 Kč 684 500 Kč 11,3 %
+6,1 %
72 833 Kč 100 826 Kč +194 Kč/den
Google Ads 1 účet 48 300 Kč 60 000 Kč 466 400 Kč 10,4 %
+5,0 %
46 000 Kč 63 000 Kč +100 Kč/den
Meta Ads 1 účet 29 000 Kč 35 000 Kč 218 100 Kč 13,3 %
+8,1 %
26 833 Kč 37 826 Kč +94 Kč/den
Total 876 500 Kč 1,18 M Kč 6,50 M Kč 13,5 % 900 833 Kč 1,14 M Kč −1 058 Kč/den
Live
Search Term Forensics · Farma Zelený Důl
Anomaly Detection · CK Velký Orient – WoW +28%
Budget Pacing · Statek Podlesí – under by −40%
Memory system

A memory that actually knows your client.

Every client has a dedicated memory block. Targets, brand terms, historical anomalies, seasonal patterns, approved decisions – every agent reads it before every run.

When it recommends raising a Monday bid because "weekends convert better on brand", that isn't a hallucination. It's an observation written two weeks ago.

L1 · Core memoryLoaded into every agent prompt – client context, targets, strategy, guardrails. Contradictions surface for your review.
L2 · Episodic memoryEvery approved / rejected change, A/B test, anomaly and your own note, vector-indexed. Your "no" weighs 3× more than automated data.
L3 · Patterns & ontologyBayes-scored rules ("tCPA under 30 conversions raises CPA by 20–40%") on a 146-concept PPC ontology with CZ+EN aliases.
L4 · Knowledge graphMemgraph with causal chains. Agents ask multi-hop questions like "what happened last Q4 when we raised brand budgets" – not just vector, but through relationships.
L5 · AutodreamMemory improves on the fly – after every action. Plus one full run a week that consolidates everything: decays stale patterns, strengthens confirmed ones, reconciles live Google Ads state with what it remembers.
L1 · CORE · always in prompt
Core Memory
CK Velký Orient – ROAS target, travel vertical. Peak Jun–Aug. Rule: do not touch brand, budget cap +20%/week.
L2 · EPISODES · 11 events · week
Episodic memory
Rejected proposal "switch to tCPA" – reason: too few conversions. Weight: .
L3 · PATTERNS · Bayes · 146 concepts
Semantic patterns
Correlation · 87% confidence – +15–20% search budget → CPA −8% within 14 days. 12 cases / 4 clients.
L4 · GRAPH · Memgraph · multi-hop
Knowledge graph
Query: budget↑ → shopping → Q4 → outcome?3 causal chains, mean ROAS impact +0.42.
L5 · AUTODREAM · after every action · Monday full run
Autodream
Revises core memory, decays patterns older than 180 d, dedupes episodes. 7 patterns reinforced, 2 deactivated.
Workflows & agents

A pipeline of specialists. Each step with its own prompt, tools and context.

Daily monitor, weekly check, monthly audit, Search Term Forensics, pMax audit, SEO/PPC arbitrage – each workflow is a sequence of agents (Data Analyst → Strategist → Auditor → Report Writer → Evaluator) that self-evaluates the output quality at the end.

active

Error Scanner

Errors, waste, tracking failures. Daily health check on every account.

daily · max 1 iter
active

Anomaly Detection

WoW metric anomalies – signal vs noise with per-vertical thresholds.

daily
active

Budget Pacing Monitor

MTD spend vs. plan, pacing %, month-end forecast within 2%.

daily · 06:00
active

Daily Change Proposals

Consolidates daily-agent output + last audit – deduplicated.

daily
active

PPC Audit

Full-account audit with health score, quick wins and roadmap.

monthly
active

pMax Audit

Performance Max deep-dive – asset groups, listings, audiences, 2-pass.

on-demand
active

Search Term Forensics

Classifies search terms: waste / high-value / brand / opportunity.

weekly
active

Weekly Review

7-day analysis and a quick written review with evaluation.

weekly
active

SEO-PPC Arbitrage

Compares GSC organic performance with paid – finds arbitrage wins.

on-demand
active

Monthly Audit

Pipeline: Data Analyst → Strategist → Auditor → Report Writer → Evaluator.

1st · 2m45s
Report & client portal

Reports your client will actually open.

Monthly reports generate on the 1st. AI drafts the narrative, you review, one click publishes. Clients get a signed URL – no login, no forwarded PDF. They can save it as an app, right to their phone home screen.

Monthly report – November 2026 published · token active
Spend (MTD)
487 320
+4.1 % vs. Oct
Revenue (MTD)
3 214 850
+12 % vs. Oct
PNO
15.2 %
−1.8pp vs. Oct
Conversions
1 847
+9.3 % vs. Oct
Spend vs. revenue – daily
Spend Revenue
Google Ads
74%
on track
Sklik
92%
on track
Meta
103%
overpace +3.5%
Spend (MTD)
487 320
+4.1 % vs. Oct
Revenue (MTD)
3 214 850
+12 % vs. Oct
PNO
15.2 %
−1.8pp vs. Oct
Conversions
1 847
+9.3 % vs. Oct
Spend vs. revenue – daily
Spend Revenue
Google Ads
74%
on track
Sklik
92%
on track
Meta
103%
overpace +3.5%
Spend (MTD)
487 320
+4.1 % vs. Oct
Revenue (MTD)
3 214 850
+12 % vs. Oct
PNO
15.2 %
−1.8pp vs. Oct
Conversions
1 847
+9.3 % vs. Oct
Spend vs. revenue – daily
Spend Revenue
Google Ads
74%
on track
Sklik
92%
on track
Meta
103%
overpace +3.5%
Real output

These are actual recommendations the agent produced.

No mock-ups. Live production output from a Google Ads account for a spa group – across 6 Performance Max campaigns, 40 asset groups, 42 channel × campaign combinations.

Unique – no other tool shows this

Performance Max channel breakdown.

The Google Ads UI does not show how much of PMAX spend goes to Search vs. YouTube vs. Discover. That data is only available via the new Ads API. Surya fetches it and splits it to campaign × channel – so you can see exactly where Smart Bidding spends.

  • Detects "YouTube waste" – channels burning budget at 0 conversions
  • Identifies campaigns where Search carries 93% of performance
  • Proposes concrete budget reallocation across campaigns
reports / pmax_channel_split.md via Ads API v16
Campaign Channel Share Spend Conv. Signal
Aurum pmax SEARCH 45 398 Kč 32.45 ✓ OK
Aurum pmax · Bellara SEARCH 30 589 Kč 23.99 ✓ OK
Aurum pmax · Hotel Pramen SEARCH 12 599 Kč 6.63 ! watch
Aurum pmax · Bellara YOUTUBE 5 681 Kč 0.00 ✗ waste
Aurum pmax · Hotel Pramen DISCOVER 1 033 Kč 0.74 ! watch
Aurum pmax · Zimní termy YOUTUBE 951 Kč 0.00 ✗ waste
42 rows total · Google Ads UI only shows the rolled-up total · exported via Ads API
CRITICAL Asset groups

25 asset groups have zero headlines – Search campaigns effectively silent.

About 62% of asset groups have no headlines at all. The Search network drives 93–99% of conversions, so these groups cannot contribute.

No headlines 25 / 40 AG
ROAS (L) 6.37
Uplift +15 – 25 %
Copywriting + a few hours → unlocks the 93% conversion channel.
WASTE YouTube · 30 d

YouTube burns 12,118 CZK/month at 0 conversions across all 6 campaigns.

8.2% of total PMAX budget goes to a channel with no measurable return.

Spend video 951 CZK
Spend non-video 370 CZK
Conversions 0
Reallocate to Search → expected CPA −15%.
WARNING Hotel Pramen

Risk of further deterioration before the fix lands.

ROAS 6.37 and CPA 2,168 CZK are structural issues. If it does not improve 30 days after the fix, that is a signal Hotel Pramen does not have enough demand for a standalone PMAX campaign.

ROAS 6.37
CPA 2,168 CZK
Conv./mo 7.4 (< 30)
Plan to merge with Aurum pmax if no improvement in 30 days.
TIMING Hotel Ambassador

Spring season starting – campaign still PAUSED.

April kicks off the main spa season. Hotel Ambassador is still paused, losing seasonal demand.

Status PAUSED
Prerequisites 3 fixes pending
Deadline 5 business days
Fix assets → reactivate before peak season.
FINDING Aurum pmax – headliner

Only campaign worth scaling. ROAS 11.00, CPA 1,407 CZK.

Best performer across the account. CTR 9.80% is well above average, Discover is close to Search CPA, asset group ad_strength is AVERAGE.

ROAS 11.00
CPA 1,407 CZK
CTR 9.80 %
Budget +20–30% + assets → defend lead into season.
run · d1e735cb-e439-4f97-b1ba · 102 / 15 000 GAds · 5.45$ Anthropic · 0 hallucinations Every recommendation is backed by Ads API data – the agent cites specific metrics, campaigns and asset groups.
Conversational layer

Ask in plain language. Across every system.

Google Ads, Sklik, Meta Ads, GA4, Search Console – plus your own data (sales, bookings, CRM). The agent reads every source and answers in one sentence, not seven open tabs.

Google Ads Sklik Meta Ads GA4 Custom data
Connected to 5 sources – incl. your custom data
01
One question.

Instead of switching between 4 UIs and exporting to Excel, you get the answer in one sentence. The agent sums, joins and contextualises across every connected account.

02
Your own data in the equation.

Upload sales, bookings, CRM export. The agent knows that "Thailand" in an ad campaign matches "TH-10d tour" in your booking system. The ontology handles the translation.

03
From question to action.

Say "cut the budget by 10%" and the agent drafts concrete changes across every related campaign. You just approve – Surya executes and logs it.

chat · session #482
How much did we spend on Thailand tour ads this month? Across all systems.
Searching "Thailand" in the ontology → match: Thailand, TH, Bangkok, Phuket, Krabi
Google Ads found 7 campaigns · 3 ad groups
Sklik found 2 campaigns
Meta Ads found 5 ad sets
Your data joined "tour_sales_2026.csv"
Total Thailand tour spend (Apr 1–17):
184,320 CZK
Google Ads
98,450 CZK
Meta Ads
54,120 CZK
Sklik
31,750 CZK
23 tours sold · ROAS 4.8 · CPA 8,014 CZK.
Cut the budget by 10%. June is sold out anyway.
Drafting change set · cross-system reallocation
Proposal – cut budget by 10% (−18,432 CZK / month)
Google Ads TH Bangkok – 10d stay 14,200 CZK 12,780 CZK −10%
Google Ads TH Phuket – all-inc 21,500 CZK 19,350 CZK −10%
Meta Ads Thailand lookalike 1% 18,000 CZK 16,200 CZK −10%
Sklik Thailand tours 10,500 CZK 9,450 CZK −10%
+ 10 more campaigns · full detail in proposal
→ follow-up questions
Ask Surya…
Change proposals

AI drafts. You review. A machine applies.

No AI at the wheel of your ad account. Every change comes as a proposal – with a precise diff, rationale and impact estimate. Approve in one click, or edit. Inspired by Git workflow.

01
Agent finds the issue and writes a proposal
02
You review the diff + rationale
03
Executor deterministically applies the merge
68
mutation operation types · Google Ads API v23
17
change categories – budget, keywords, PMAX, audience…
6
safety layers · complete audit trail
1
click to rollback · pre-change state preserved
Safety

6 layers every proposal must pass.

Each step is enforced automatically. The agent cannot bypass them – they are part of the execution pipeline, not the system prompt.

01

Account Isolation

Every proposal carries org_id + ad_account_id. The executor verifies the account actually belongs to that organisation. No cross-tenant access.

02

Client Change Rules

Per-client configurable rules: "Never pause Brand Search" or "No more than +30% budget without extra confirmation". The agent receives them in context.

03

System Guardrails

Automated safety checks: budget >100% → warning, >200% → blocked. Wiping all keywords from an ad group → blocked. Hard-coded, not an LLM call.

04

Dry-run validation

Google Ads API `validate_only: true` – at proposal creation AND right before execution. Double validation, because account state can drift in between.

05

Human approval

No change without explicit approval. You see: current state → proposed state, rationale, estimated impact, who wrote it.

06

Rollback

Before every execution we store the full pre-change state. One-click "Undo change" any time afterward. Plus a full audit trail.

Onboarding

From signup to first report in one afternoon.

01

Connect accounts

OAuth into Google Ads, Sklik, Meta, GA4, Search Console, Merchant Center. 90 days of backfill in 10 minutes.

02

Set targets

Per client: ROAS, brand terms, hourly rate. Agents pull it into memory automatically.

03

Approve first change

You'll get daily proposals. Click approve – Surya executes and logs it. Manual work is optional.

5×
memory layers – Core, Episodic, Patterns, Graph, Autodream
146
PPC concepts in the ontology (CZ + EN aliases)
20yr
of agency PPC experience baked into the system
10
workflows across daily, weekly and monthly cadence
Social proof

Our own agency. 20+ clients. €120k+ monthly spend.

Surya is not an unproven alpha. At our agency Surya Marketing we run 20+ clients in the app with combined media spend north of €120,000 every month. The app has been running in production for several months — hardened on the edge cases an ordinary test never sees: broken conversion tags, mid-day budget reshuffles, pause/resume across time zones, Black Friday peaks. It works. Every day.

20+
Surya Marketing clients in the app
€120k+
monthly media spend under management
300+
campaigns under daily ops
99.95%
uptime over the last 90 days
Case study · production run Spa group · 6 hotels·Google Ads · Performance Max·November 2025 – ongoing

From a 12-hour monthly audit to 90 minutes of review.

The PPC manager got a full account audit across 6 PMAX campaigns, 40 asset groups and 42 channel × campaign combinations — straight from Ads API, not Google Ads UI screenshots. Surya identified 3 critical fixes, 2 waste leaks and 1 campaign worth scaling. Each with concrete numbers and a recommended action.

11h/mo
saved on monthly audit
baseline: ~12h manual
12,118 CZK
wasted spend identified on YouTube
0 conversions, 30 days
+15–25%
expected uplift after fixing 25 AGs
93% conversions via Search
  • Conversion action with includeInConversions=FALSE — Smart Bidding was optimizing for the wrong goal across 6 PMAX campaigns.
  • 25 of 40 asset groups had no headlines — the Search network, which drives 93–99% of conversions, had no creative.
  • Aurum pmax the only campaign worth scaling: ROAS 11.00, CPA 1,407 CZK, CTR 9.80%.
Full case study on request · NDA client Request details →
Why not manual

What Surya actually adds to what you do today.

Most PPC agencies today run clients with a mix of Google Ads UI, a shared spreadsheet and a monthly Looker Studio report. It works — until you have more than 5 clients. Here's where Surya moves the work elsewhere.

What Surya solves
Surya Surya Multi-agent platform
Status quo Manual Google Ads + Sheets + Looker
Monitoring coverage Who watches the account between review meetings.
10 specialist agents, daily/weekly cadence
The manager has to remember to check
Multi-agent vs. one chatbot It's not ChatGPT with a prompt engineer.
Specialist pipeline: Data Analyst → Strategist → Auditor → Report Writer → Evaluator. Each has one job.
A generic "AI assistant" in Excel or a Sheets plugin won't produce structured output.
Memory across clients 5 layers. The agent remembers what worked where and when.
Core, Episodic, Patterns, Knowledge Graph, Autodream — the agent recommends from history, not from a blank page.
The manager remembers what worked. When they leave, memory leaves with them.
PMAX channel split How much of PMAX budget goes to Search vs. YouTube vs. Discover.
Via Ads API v16, broken down to campaign × channel. Detects YouTube waste.
The Google Ads UI doesn't show this. Without your own API client, you can't reach it.
Monthly client report From data to a published report.
Auto-generated on the 1st, AI writes the narrative, you review, one click publishes. Client opens a signed URL — no login.
Looker Studio + a manual write-up. 2–4 hours per client per month.
Change approval Who controls the account.
One-click approval. Full audit trail, rollback. Full autonomy is opt-in per client.
Manual edits in Google Ads UI or Editor.
Local support and integrations Sklik, Heureka, Zboží.cz — not just the US market.
Sklik, Heureka, Zboží natively integrated. Czech UI. Czech onboarding.
Up to you — but it's your work.
This comparison does not include point tools like Optmyzr or Adalysis — Surya doesn't directly compete with them, it solves a different problem (orchestration across clients, not rules within one account).
Pricing

Same features for everyone. You only pay by client count.

Everyone gets everything – 10 workflows, 5-layer memory, AI chat over all data, client portal. Migration included.

In every plan:
  • All 10 workflows
  • 5-layer agent memory
  • AI chat over all data
  • Daily change proposals
  • Monthly reports with publishing
  • Client portal (link-only)
  • A/B tests with auto-evaluation
  • Autodream – continuous + weekly consolidation
  • Knowledge graph & ontology
  • Slack + email alerts
  • Billing integration
  • All integrations (GAds, Sklik, Meta, GA4, SC, Merchant)
What differs:

Solo

1 client

2 990 Kč / month
  • 1 client
  • up to 3 seats
  • All features – no gating

Starter

up to 5 clients

6 990 Kč / month
  • up to 5 clients
  • up to 10 seats
  • All features – no gating

Business

up to 20 clients

25 990 Kč / month
  • up to 20 clients
  • unlimited seats
  • All features – no gating
Enterprise

An agency with its own rules?

For agencies with 30+ clients, in-house teams, and white-label setups. Custom guardrails, SLA, onboarding with our team, integration to your ERP/billing.

Talk to us
  • Unlimited clients
  • White-label client portal
  • Dedicated onboarding (20 yrs of PPC)
  • Custom knowledge base & playbooks
  • 99.9% SLA · single-tenant available
  • Direct Slack channel to our team
Security & Compliance

Security and compliance you can count on.

Surya is built to meet the highest standards for protecting your clients' data and your agency. Transparently, without marketing fluff.

GDPR

GDPR compliant

We meet the requirements. DPA available on request.

Live
EU

Hosted in the EU

Primary infra in the EU (Railway EU region). Leaves only for AI inference.

Live
DATA

EU Data Act

Your data is yours. One-click export any time – frictionless migration.

Live
SSO

SSO (SAML · OAuth)

In preparation for Enterprise. 2FA and role-based access today.

Coming
AUDIT

Audit trail

Every action, every approved change, every report – full history.

Live
+

Runs on the best AI – and it does not learn on your data.

Reasoning and analysis are powered by Anthropic Claude – the best model for strategic reasoning and agent orchestration according to independent benchmarks and our own tests. We use OpenAI for vector search (embeddings).

Zero retention Anthropic API does not use customer data to train models. Standard 30-day retention can be turned off on Enterprise plans.
SOC 2 Type II Anthropic holds SOC 2 Type II and ISO 27001. All API calls are encrypted end-to-end.
No third-party sharing We do not sell, share, or use client data for benchmarks or marketing. Only for the functionality you pay for.

Where does your data go?

Transparent – so you know exactly what leaves your infrastructure and why.

Stays with Surya

  • Login credentials (encrypted)
  • OAuth tokens (Google Ads / Meta)
  • Raw campaign data and reports
  • Action history and audit trail
  • Uploaded custom data (CSV, sales)

Anthropic (Claude)

  • Aggregated metrics for analysis
  • Prompt context (what the agent needs)
  • Memory contents (episodes, patterns)
  • Zero retention · no training

OpenAI (embeddings only)

  • Short text snippets for vectorisation
  • No full reports, no credentials
  • Just for memory indexing
  • No training on customer data
Your data stays yours. Surya is hosted in the EU (Frankfurt), client data is logically isolated per workspace, encrypted at rest (AES-256) and in transit (TLS 1.3). No one on the Surya team can read account contents without your explicit access request (with an audit log). We do not train models on your data.
By default, you approve. Every change is a proposal with reasoning. You click approve and Surya executes via Google Ads API. Full autonomy (auto-execute for selected change types — pause low-perf keyword, budget reallocation within a cap) is opt-in per workspace and per action type. Always with a full audit trail and one-click rollback.
Three layers: (1) Evaluator agent reviews output before it reaches you. (2) If you approve a mistake, rollback is one click in the audit log — Surya reverts to the prior state. (3) Workspace-level guardrails (max budget delta per day, keyword blacklist) cover systemic fail-cases.
OAuth account connection: 10 minutes. First full audit across clients: 2–4 hours (depends on campaign count). First monthly report in your template: 1 week (we tune the tone-of-voice together). Migrating an existing Excel/Sheets setup: free, the onboarding team does it.
Yes, anytime. Full export of agent memory, all reports, audit log and client configurations as JSON + PDF. After you request deletion, we erase data within 30 days (GDPR Article 17). No vendor lock-in — your Google Ads / Sklik / Meta connections are yours, we don't do anything you couldn't cancel directly in the platform.
Yes. Sklik is integrated natively via Sklik API (campaigns, ad groups, keywords, reports). Heureka and Zboží.cz via partner APIs — feed audit, bidding, performance breakdown. Default setup for the CZ market. Meta and LinkedIn are in beta — if your mix is paid-social heavy, mention it during onboarding.
All 10 workflows run on every client (Error Scanner, Anomaly Detection, Budget Pacing, Daily Proposals, PPC Audit, pMax Audit, Search Term Forensics, Weekly Review, SEO-PPC Arbitrage, Monthly Audit). Each workflow is a pipeline of 3–5 specialist agents. Cadence is configured per client (daily / weekly / monthly / on-demand).
On Pro and Enterprise plans, yes. No-code custom workflow builders are on the roadmap for Q3 2026. Today: your specific need (e.g. "weekly competitor monitoring via SEMrush") gets built as a custom agent during onboarding. Typically 3–5 business days.
No hard stop, no surprise invoices. After you go over, you get an email proposing an upgrade. If you don't respond, we keep running on your current plan — overflow clients move to read-only mode (you can't add new ones, but the existing ones keep working). Upgrade anytime, prorated.
Fully bilingual. UI, agent output, reports and FAQ are CZ + EN. The 146-concept PPC ontology has aliases in both languages — the agent understands "CTR" and "míra prokliku" alike. If you have international clients, one client can have a CZ team and EN reports at the same time.
Contact

Book a demo or write us.

Tell us a few sentences about how many clients you run and what hurts most about your current setup. We respond within one business day with a next step — either a 30-min demo, or straight into onboarding.

Email info@suryatool.io
We respond within 24 hours on business days.
Office Korunní 2569/108, 10100 Prague 10
Engineering team in Prague.
Demo 30 minutes, online
We'll show output on your data, not mockups.
Book a demo Fields marked with * are required
Thanks — we'll be in touch.
We've received your message. You'll hear from us within one business day at the email you provided.
Helps us tailor the demo.
We don't share your data. GDPR-compliant.

Stop losing Fridays to Google Ads exports.

Try Surya free for 14 days. No card, full client migration included.

Start now Book a demo