Talk to your ads.
In plain English.
Ott's MCP endpoint connects your PPC campaigns directly to AI agents — Claude, ChatGPT, Gemini, and any tool that speaks the Model Context Protocol. Twenty-six analytics tools exposed through a single endpoint. Ask “which forex campaigns are approaching budget overdraft?” and get an answer pulled from live Meta data. Compliance-aware schema. Telegram attribution built in. Agency hierarchy respected.
Stop exporting CSVs.
Start asking questions.
No rate limits. No API keys.
Just answers.
Ott syncs your data daily. MCP queries hit our indexed layer — so you skip the Meta Marketing API entirely. Fast, reliable, and never throttled.
No Meta rate limits
Ott syncs your data daily via read-only OAuth. Your MCP queries hit our indexed data, not the Meta API directly — so you never hit rate limits or wait for throttled responses.
All your data, every data type
Campaigns, ad sets, ads, spend, impressions, CTR, CPM, CPC, conversions, Telegram joins, FTDs, deposits — everything Ott tracks is queryable through MCP.
Works with any AI
Claude Desktop, ChatGPT, Cursor, Continue.dev, Cody — any client that speaks the Model Context Protocol. One endpoint, universal access.
No API keys to manage
Your AI client connects to Ott, not to Meta. One secure OAuth setup. No access tokens to rotate, no permissions to debug, no accidental write access to your ad accounts.
One line of config
Add the Ott MCP endpoint to your client config. That is it. No SDK, no library, no NPM package. The protocol handles everything.
Ask in plain English
"Which campaigns went over budget this week?" "Show me declining creatives." "Compare CPA across Telegram and website conversions." Natural language, real answers.
Compliance-aware schema
Every data point includes policy classification metadata and audit trail context — so your AI agent can answer "which campaigns ran under restricted financial categories?" from live data.
Telegram attribution built in
Per-ad, per-creative Telegram join tracking via CAPI postback. Ask "which creative has the lowest CPJ this week?" and get answers from real join data — not pixel approximations.
Agency hierarchy respected
Client → Brand → Account hierarchy built into every response. Ask about "all forex campaigns across client X" and Ott returns data organized the way your agency actually works.
The MCP land grab is happening.
But nobody's building for finance.
In the last three months, the PPC analytics industry shipped a wave of MCP endpoints. All of them answer 'give me my campaign data in a chat window.' None of them answer 'can you prove this campaign was compliant with financial advertising policy?'
| Company | MCP Launch | What It Connects | Finance-Aware? |
|---|---|---|---|
| Madgicx | March 2026 | Meta Ads only | No — ecommerce only |
| Supermetrics | June 2026 | Multi-platform via Claude | No — generic data pipeline |
| Measured | June 2026 | Incrementality data | No — brand analytics |
| Synter Media | Early 2026 | 20+ platforms, AI agent operator | No — horizontal platform |
| Windsor.ai | June 2026 | Live-writing (campaign control) | No — generic MCP |
| AdRoll | June 2026 | Campaign management via text | No — SMB-focused |
| Google (open-source) | June 2026 | Ads API MCP reference implementation | No — raw API proxy |
| Ott | Now | Meta Ads + 26 tools, Telegram attribution, compliance layer | Yes — built for regulated verticals |
Seven MCP endpoints in a single quarter. Only one understands compliance, Telegram attribution, and agency hierarchy — the data layers that finance verticals actually need.
Connect once.
Ask anything.
Setting up Ott's MCP endpoint takes two minutes — and your AI agent keeps the connection permanently.
Generate your API key
In Ott, go to Settings → API Keys. Create a key scoped to your workspace. Copy it once. Ott never shows it again.
Configure your AI agent
Add so.ott/ppc-analytics to your MCP client — Claude Desktop, ChatGPT, Cursor, or any MCP-compatible tool. Paste your API key and workspace ID as configuration variables. That's it.
Ask your first question
Open your AI agent. Type a natural-language question about your campaigns. Ott returns structured data with compliance context, agency hierarchy, and the metrics that matter for finance verticals.
What happens under the hood
- Your AI agent sends a query through the MCP protocol (Streamable HTTP transport)
- Ott authenticates via API key + workspace ID
- Ott queries the Meta Marketing API for live campaign data
- Ott enriches the response with compliance metadata, Telegram attribution, and hierarchy context
- Your AI agent receives the data and translates it into a natural-language answer
No CSV exports. No spreadsheet formulas. No stale data. Live campaign intelligence, on demand.
Technical specs
| Spec | Detail |
|---|---|
| Protocol | Model Context Protocol (MCP) |
| Transport | Streamable HTTP |
| Endpoint | https://api.ott.so/mcp |
| Namespace | so.ott/ppc-analytics (registered in MCP Registry) |
| Tools | 26 read-only tools |
| Authentication | API key + workspace ID |
| Data freshness | Real-time (sourced from Meta Marketing API) |
| Compatibility | Claude Desktop, ChatGPT, Cursor, any MCP-compatible client |
When Meta bans an ad account,
your AI agent should have answers.
Meta banned 1,000 illegal financial ads in a single week in the UK alone. When an account ban hits your client, the conversation doesn't start with 'what was our ROAS?' It starts with 'can you prove we were compliant?'
Policy classification metadata
Which campaigns targeted which regulated categories. When they ran. Under what policy versions. Every campaign carries its classification history — queryable through MCP.
Activity audit trail
Every creative change, every budget adjustment, every campaign toggle. Who made it. When. Why (if noted). Full traceability for every data point.
Telegram conversion paths
Per-ad, per-creative attribution for Telegram joins via CAPI postback. Not generic "conversions." Actual member joins with source traceability.
Hierarchy context
Which client, which brand, which Business Manager, which ad account. Full lineage for every data point — so your audit trail tells the complete story.
Try it: Ask your AI agent — “Show me all policy-classified campaigns for Client X that ran in May — and pull the activity log for any creative changes made to them.” Your agent queries Ott's MCP. Ott returns the campaigns, their policy classifications, and the full audit trail. Your agent presents a compliance-ready summary.
That's the difference between a generic MCP endpoint and one built for regulated verticals.
What will you ask
your ads today?
These are real queries that media buyers and agency operators run every day through Ott MCP. Copy one and try it.
“Which campaigns went over budget?”
Ask once. Ott returns every campaign across every account that exceeded its daily or lifetime cap — with the exact amount, the date, and the brand it belongs to.
“Show me declining creatives”
Ask which creatives have dropped more than 20% in CTR over the past week. Get a ranked list with before/after metrics. Spot the tired ads before they burn budget.
“Compare CPM across client groups”
"What is the average CPM for Forex clients vs Crypto clients this month?" Cross-account, cross-brand comparison in one natural-language query.
“Which Telegram channels have the best CPA?”
Ott tracks joins per channel. Ask MCP which channels deliver the lowest cost-per-join, then optimize toward the winner via CAPI.
“Pull the compliance audit trail”
"Show me all policy-classified campaigns for Client X in Q2, with the activity log for any changes made." Compliance-ready summary in one query.
Your analytics should speak
the language of your clients.
Finance agencies — forex, crypto, iGaming, prop firms — don't need another ecommerce analytics tool that happens to show Meta data. They need analytics that understand regulated verticals.
The metrics finance clients actually care about
- Cost per Telegram Joinee (CPJ) — not generic CPA
- FTD tracking (First Time Deposit) — per campaign, per creative
- Thumbstop rate — because Telegram ads live and die by the hook
- Multi-currency spend — USD, EUR, GBP, USDT per brand
The compliance reality
- Meta's financial services advertising policies change. Frequently.
- Account bans happen without warning.
- When they do, your client needs proof — not guesses.
- Ott's audit trail gives you the evidence before you need it.
The agency structure
- Client → Brand → Business Manager → Ad Account
- Prepaid budget ledgers per client, not per campaign
- Private Mode for hiding agency margin from client views
- Full hierarchy respected in every MCP response
Generic MCP endpoints return generic Meta metrics. Ott's MCP returns finance-native analytics — with the hierarchy, compliance, and attribution layers that regulated verticals actually require.
Questions you might have
about MCP.
Is this just a chatbot wrapped around Meta Ads Manager?▼
Does my AI agent need special training to use Ott's MCP?▼
Can my AI agent make changes to campaigns through MCP?▼
How is this different from Supermetrics for Claude?▼
What about Synter Media's "One MCP for all media buying"?▼
Is my campaign data safe with an AI agent?▼
Works with any MCP client.
One endpoint. Every AI tool that speaks the Model Context Protocol. No per-client configuration.
Claude Desktop · ChatGPT · Cursor · Continue.dev · Cody · Windsurf · Any MCP-compatible client
Related features and comparisons
Track Telegram joins per ad and creative via CAPI postback. The attribution finance agencies actually optimize for.
Purpose-built for regulated verticals: multi-currency budgets, compliance audit trails, agency hierarchy.
Compare thumbstop rates, CPJ, and fatigue signals across all creatives — ask MCP to rank them by performance.
Multi-currency prepaid ledgers with overdraft alerts. Ask MCP "which clients are within 10% of budget cap?"
Morning campaign check: flag anomalies, surface underperformers, and spot spend outliers across every account in seconds.
Your campaigns should answer questions,
not generate exports.
The MCP era of marketing analytics is here. Seven competitors shipped MCP endpoints this quarter. Only one understands regulated verticals — Telegram attribution, FTD tracking, agency hierarchy, multi-currency budgets, compliance audit trails. Your AI agent is ready to work. Give it the analytics stack that speaks your industry.