PlatPhorm Podcasts
Privacy
Public playback, search, and browsing do not require sign-in. PlatPhorm Podcasts minimizes collection while keeping source attribution and service reliability visible.
Information we process
Public source URLs, normalized media metadata, attribution, votes, flags, request identifiers, coarse runtime diagnostics, and abuse-prevention signals may be processed to operate the index. We do not require an account for ordinary public use. Follow, play, useful, playlist, and freshness actions are related to public content records, not to a user profile.
Browser data and cookies
A draft queue, player preferences, and bounded offline cache may remain in your browser. We do not use advertising cookies, paid-placement tracking, or covert cross-site identity. Original publishers may apply their own policies when their media is requested.
Storage, retention, and sharing
Cloudflare D1 stores canonical source identity, metadata, provenance, community records, and operational evidence. KV contains derived or short-lived cache state, and R2 is limited to bounded artifacts such as permitted snapshots or analysis outputs. Media is normally streamed on demand from its attributed source rather than copied permanently. Action deduplication uses a one-way, short-lived scope for system integrity; it is not a persistent user fingerprint, cross-site identity, or advertising identifier. Public surfaces expose aggregate content signals, while raw action and request details remain bounded and protected. Operational records are retained only as needed for reliability, security, abuse mitigation, and auditability; secrets, authorization headers, cookies, raw IPs, and raw private credentials are not published.
Locality and language
Coarse country, region, city, and timezone hints supplied by the request platform may improve live and nearby discovery without storing a raw IP address. Precise browser location is opt-in, is used only after an explicit request, and is not required for playback. UI locale and preferred content languages are separate browser preferences so changing interface language does not create a location or identity profile.
Automation and AI
Queues and Workflows perform bounded ingestion, health, repair, enrichment, and ranking work. AI may interpret a request, expand a query, summarize, classify, or rank ambiguous repair candidates, but it is not evidence for playability, HTTP status, codec, distance, licensing, vote counts, or canonical identity. Public results remain grounded in attributed source or persisted evidence, with deterministic fallback when AI is unavailable.
Security and your choices
You may clear browser-local data through your browser, use the original publisher directly, or request correction of inaccurate attribution and public metadata. We apply bounded URL validation, redaction, and rate limits to protect the service.
Contact and support
For a privacy question, source correction, or removal request, use the contact page or support page and include the public source URL. Do not send API keys, cookies, or private credentials.