Your podcast numbers don't agree with each other. That isn't your fault.
Apple, Spotify, YouTube, and Amazon — collected by your own Claude, normalized to one schema, with every definition documented.
Every platform counts differently.
None of them will reconcile for you.
Podcasters are asked to run professional media businesses on dashboards that measure incompatible phenomena and silently rewrite historical baselines.
The numbers don't reconcile.
A Spotify play, an Apple play, a YouTube view, and an Amazon play count four fundamentally different physical actions. Adding them together in a spreadsheet produces a total that means mathematically nothing.
Your history breaks silently.
iOS 17 wiped ~15% of downloads industry-wide with zero change in actual listening behavior. Spotify redefined "play" on June 11, 2026 and never restated historical baseline numbers. Nobody emailed you.
"We thought our show was dying. Turns out Apple simply stopped downloading queued episodes to paused devices."
Exports rot.
Dashboards alter their schemas, CSV exports break formulas, and the history you need for a sponsor audit is gone because platforms keep exports for a limited window and restate the past without notice.
A single, immutable truth layer for your show.
Collection is just the plumbing. Cross-platform normalization with documented methodology is the product.
Your Claude Collects. We Never Log In.
A Skill runs in your own Claude and drives your own browser, signed in as you, to pull the exports from Spotify, Apple, YouTube, and Amazon and push them to the connector. About ten minutes a week. Your platform passwords never touch us.
One Normalized Schema
Every export lands in one schema, each metric tagged with the platform's own definition and the caveats that go with it. Nothing is blended silently; where we estimate, the estimate carries its coverage and its bounds.
Per-Episode Retention Curves
Drop-off curves from the platforms that publish them, per episode, side by side. See the minute listeners leave and whether a sponsor read is where it happens.
Coverage You Can Audit
Every answer names the platforms it touches, the periods each one covers, and the gaps. A coverage report shows exactly what has been collected and what hasn't, so a number is never mistaken for a complete one.
Permanent Historical Vault
Platforms keep exports for a limited window. Once collected, your observations stay in your workspace, each one tied to the raw file it came from, so any figure can be traced back to its source.
The MCP Connector
There is no dashboard. You add the connector to your own Claude and ask in plain language; every answer comes back with the numbers, the definitions behind them, and the caveats the platforms bury.
A simulated Model Context Protocol session with illustrative numbers — click a prompt to see the shape of an answer:
"We don't just give you numbers.
We tell you what they mean and when they changed."
Trust is earned through absolute transparency. Here is how each primary network calculates listening activity — and how our normalization engine harmonizes them.
| Platform | Unit Name | Current Measurement Rule | Previous Historical Rule | Normalization Action |
|---|---|---|---|---|
| Spotify | Play (Stream) |
≥ 30 seconds continuous stream.
Effective June 11, 2026.
|
≥ 60 seconds | Kept in Spotify's own unit with the effective date attached. Never summed with another platform's plays; the rule change rides your chart as an annotation. |
| Apple Podcasts | Play / Event | 0 seconds minimum. Any user initiation tap counts as 1 play. | Pre-iOS 17: Automatic downloads counted toward host metrics. | Kept as Apple plays with Apple's definition attached, shown beside Apple's own listening-time and follower exports rather than converted into them. |
| Amazon Music | Unique Listener Play | ≥ 60 seconds, strictly deduplicated once per listener per 24h day. | Unchanged since 2022 | Kept as Amazon's unique-listener unit and labeled as such. Shown beside the others as per-platform reach, never folded into a deduplicated total. |
| YouTube | View | Counted at playback start under YouTube's 2026 update. | ~30 seconds intent threshold | Kept as YouTube views with the definition in force on each date; watch time is reported alongside, not converted into plays. |
Want the full list of definitions before you join?
Every metric we serve carries its platform's definition and our caveats. Ask and we'll send the current definitions list.
Annotations that ride your charts. Few analytics tools do this.
When your numbers drop or spike, your immediate fear is content fatigue or host mistakes. In reality, platform engineering teams made an unannounced algorithmic or counting shift.
intelligence.fm keeps a ledger of industry events and platform definition changes and lays them onto your timeline, and you can add your own. You'll never waste hours investigating a "problem" that happened to the whole industry.
Phase 1 Shadow Reporting is live.
Why Now?
The Accords established the first cross-industry definition of an authentic play. Yet no platform reports it natively. We estimate it and blend it — with published error bars.
Confidence band: [257,395 — 272,225 genuine listening plays]
How we access your data.
Zero ambiguity. Zero pooling.
You never hand us dashboard access at all. Here is exactly how collection works, and what we keep.
Read-Only By Architecture
Collection uses your dashboards' own export buttons, in your browser. We hold no credential that could publish, edit an episode, change a feed, or delete anything.
No Passwords. Ever.
Your Claude drives your own browser, signed in as you. We never ask for, store, or see a platform login. Where a platform offers an API credential, it stays in your own keychain; our server records status only.
Zero Data Reselling
Your numbers are yours alone. We do not sell analytics, listener metadata, or attribution reports to third-party ad brokers.
No Unconsented Pooling
We never blend your show's numbers into industry benchmarks without your explicit opt-in. Every question you ask is recorded in a log your workspace can see, and the privacy policy lists exactly what we keep.
Built for podcasters who treat audio as a real craft.
Framed by operational need rather than pricing tiers.
Independent Creators
Solo hosts and indie teams managing 1 to 2 shows.
- Stop stitching 4 different CSV downloads every single month.
- Understand exactly why your numbers dropped before panicking.
- Query your retention and listener trends directly in Claude.
Studios & Agencies
Production shops managing multi-client rosters (3–15 feeds).
- Produce sponsor reporting decks backed by documented methodology.
- Compare drop-off rates across clients to diagnose editorial pacing.
- One workspace per client, with reports that name every platform and caveat.
Podcast Networks
Large media operations and catalog owners (15+ feeds).
- AMP Accords Phase 1 shadow estimates, with coverage stated on every figure.
- Many shows in one workspace, with the same definitions across the whole catalog.
- The connector in your team's own Claude. No dashboard seats to buy.
What's live vs. what's coming.
We believe in honest roadmaps. Knowing what we haven't built yet is why our waitlist represents real early access, not an empty newsletter.
Core Foundation
- ✓ Collection Skills for Spotify, Apple, YouTube, Amazon
- ✓ API routes for Apple, Spotify, YouTube where you hold the credential
- ✓ One schema; definitions and caveats on every answer
- ✓ The MCP connector in your own Claude
- ✓ Per-episode retention curves
- ✓ AMP Phase 1 shadow estimates with coverage
- ✓ Industry annotations and a coverage report
In Progress
- ⌛ Weekly digest and collection reminders
- ⌛ Self-serve export and delete of your own data
- ⌛ Alerts when a platform changes a definition
Beyond the Beta
- ○ AMP Phase 1 alignment as the industry finalizes it
- ○ Agency workspaces with per-show access
- ○ Chart ranks and reviews
- ○ Pricing, billed per show
Join the Private Beta.
We onboard creators in weekly batches based on RSS feed coverage and platform diversity. Fill out the four fields below.
You're on the list.
We've sent a confirmation to your email. We onboard podcasters in small batches and will email you when yours opens.
Questions in the meantime? Reply to the confirmation email and a human answers.
Everything you need to know.
Honest answers about data sovereignty, platform coverage, and the beta timeline.
Stop guessing why your podcast numbers changed.
Get the ground truth. Join our private beta waitlist today.
Join the Beta Waitlist