Studio Integrations
Connect Slack, Klaviyo, Mailchimp, Prolific, CSV, and MCP to notify your team and grade Studio predictions against the real results in your own tools.
Integrations connect Studio to the tools where your real results live. They do two jobs: push a verdict to your team, and pull real outcomes back so Studio can grade its locked predictions. Open them from the Integrations button on the left rail of any project.
The spend rule
One rule sits at the top of the Integrations drawer and never moves:
"Hard rule: ad spend always happens in your own accounts. iMario never holds or moves your media budget."
Studio can read your results and, for Prolific, launch a panel on your account, but it never touches your media budget.
What connects today
Each integration shows a status: ✓ Connected, … Pending review, Not connected, or Coming soon.
| Integration | What it does | Availability |
|---|---|---|
| CSV backfill | Grade a prediction from exported results (variant, impressions, conversions) | Always available |
| Email notifications | Run-completion and day-3 and day-7 grading reminders | Always on |
| Slack | Post run verdicts and ledger grades to a channel | Connect an incoming webhook |
| Klaviyo | Read email A/B results to grade locked predictions | Connect a private API key |
| Mailchimp | Read campaign A/B results to grade locked predictions | Connect an API key |
| Prolific | Launch a real human panel when the reviewers abstain | Connect an API token |
| Meta Ads (read-only) | Read your A/B results to grade predictions | Pending Meta App Review; demo mode available |
| MCP server (generic) | Read results from any Model Context Protocol server | Connect a server URL |
| Meta Ads MCP | Read Meta ads results over MCP (write stays paused) | Connect an endpoint |
Coming soon
The registry also lists integrations that are not built yet and are marked Coming soon: Meta Ads (launch), Stripe MCP, TikTok Ads, and Google Ads. Reading results comes first; launching spend, where it lands, always happens on your own account.
Grading with real results
Klaviyo, Mailchimp, CSV, and Prolific all feed the same accountability loop: they bring real outcomes back so Studio can grade a locked prediction. See Grade your predictions for how a graded result is scored and recorded.
@imario in your own tools
Studio also exposes a reaction probe you can call from your own tools through the Model Context Protocol, under the name @imario. It returns friction points and objections, and it is deliberate about what it is not:
"Directional reaction probe, not an outcome prediction. Use it to catch objections and friction before you ship, not to decide which version will win."
Use it the same way you use a check: to catch problems early, never to predict which version will win.
Next steps
- Grade your predictions: connect these results back to your locked verdicts.
- Read results and verdicts: where a verdict is created and shared before it is graded.
- How content checks work: why Studio catches objections instead of predicting performance.
Grade Your Predictions
Studio locks every verdict, then grades itself against your real numbers with two figures, so you can calibrate how much to trust each check.
Test an Ad or Copy
Screen ads, headlines, emails, and messaging against a synthetic version of your target reader before you spend, and see exactly why copy lands or misses.