PodPitch

What PodPitch does

PodPitch provides two distinct workflows through MCP. Public discovery lets an agent search active podcasts by topic or name and inspect the resulting public show records without a PodPitch account. Results include public information and listening links, but not contact email addresses.

Existing PodPitch customers can connect an account in the browser and use their team context to manage podcast campaigns. Account-backed operations include profiles, campaign settings, saved pitch drafts, previews, audits, subject-line results, scheduling, outreach, and inbox conversations. The available actions remain subject to the connected team's permissions, subscription rules, campaign mode, plan checks, and pitch limits.

The server also includes a bug-reporting tool for recording a minimal, reproducible PodPitch tool failure without an account. Reports return a persisted report ID and should not contain credentials, personal data, pitches, or conversations.

How it works

The MCP endpoint is https://api.podpitch.com/v1/mcp. Public search and detail tools can be used without authentication. Authenticated functions use the PodPitch account represented by the browser connection, including its stable opaque profile ID, display email, team profiles, and connected senders.

Campaign operations cover creation, configuration, mode changes, draft editing, contact selection, sending, scheduling, and schedule control. Draft and list operations use pagination where applicable: campaign lists use next_before_id, draft results use next_offset, and inbox threads provide pagination. After queued draft generation, read the saved content and is_regenerating state to confirm completion rather than treating the acknowledgement as final content.

A Claude package also contains a dated, sharded catalog for hosts that cannot reach live discovery. It supports topic and category searches, exact show details, audience estimates, and demo links from the bundled records when the host can access the package files and run its helper.

Setup and configuration

Connect an existing PodPitch account in the browser before using account-scoped tools. Public podcast discovery does not require an account. MCP clients connect through the published endpoint, while the repository also identifies ChatGPT and Claude packages for their respective environments.

PodPitch applies the connected team's existing authorization and plan checks. Campaign creation starts in Off mode, and scheduling is handled separately through schedule_campaign. Follow the client's approval rules before sending pitches, replying to inbox threads, or reporting failures.

Tools and capabilities

• Search active podcasts and read public or account-accessible podcast records.
• List, create, archive, and restore team profiles.
• Create campaigns and update writing, follow-up, tracking, and mode settings.
• Read campaign settings, drafts, previews, audits, schedules, and subject-line results.
• Add contacts, send pitches, schedule campaigns, and pause or disable pending sends.
• Search inbox threads, read conversations, and send replies.
• Record a minimal technical failure report and receive its report ID.

Limitations and notes

Public discovery intentionally excludes contact addresses. Account operations require an existing connected PodPitch customer account and remain constrained by team permissions, subscription status, campaign state, and daily or plan-based pitch limits. MagicPitch does not allow Manual campaign mode. Editing is stage-sensitive: refresh before editing a saved unscheduled pitch, and use the upcoming-send editor for scheduled messages.

The Claude catalog is a dated snapshot rather than live discovery. Its records and package checks support offline-compatible lookup, but host permissions still control whether the bundled files and helper can be used.