Taking notes on podcasts with Snipd, Readwise, and Obsidian

Snipd turns podcast listening into a skimmable, note-ready workflow by combining AI transcription, automatic chapters, and “snips” that export cleanly into Readwise and then Obsidian. The payoff is practical: instead of committing to an entire audio episode, listeners can jump to relevant sections, capture highlights in seconds, and later review everything with linked notes and AI-generated takeaways.

The core experience starts with mobile podcast playback on iOS and Android, where episodes can be transcribed in English and segmented into chapters—functioning like YouTube-style timestamps. For shows that haven’t been processed yet, Snipd can request AI features and return them after roughly 15 to 30 minutes. Once enabled, the transcript is interactive: while audio plays, words highlight in sync, and chapters make it easy to skip directly to the parts that matter.

Capturing notes is built around “snips,” which are essentially user-defined moments inside a transcript. A listener taps to create a snip, then can edit the exact boundaries of what gets captured, rename it, add a private note, and optionally tag it. Snipd also surfaces social proof by showing the most snipped highlights from other listeners—similar in spirit to Kindle’s highlighted sections—so users can quickly see what others found valuable and decide whether to continue the episode or revisit a specific segment.

A newer beta feature, “AI podcast notes,” adds another layer of automation. After marking an episode as completed, beta users receive an email containing a link that generates takeaways (six in the example shown) without requiring manual summarization. The link is portable: it can play back the relevant snipped portion and display the transcript, the two-minute snip details, and episode-level notes. Access is gated via a wait list, but an invite code (“Nicole”) is offered to jump ahead.

The second major reason Snipd fits into a knowledge-management stack is integration. Although Snipd can export highlights as Markdown for tools like Obsidian or Bear, its strongest path is through Readwise. Readwise’s plugin then brings Snipd content into Obsidian, preserving structure: updated titles, key takeaways tied to each highlight, and—when available—automatic linking to existing notes (for example, a “YouTube” note). This reduces the manual linking work that often breaks the flow from consumption to writing.

Still, there are clear limitations. Snipd is English-only, which pushes users toward English podcasts. Fiction-heavy shows—especially narrative formats like Dungeons & Dragons actual plays—don’t always map cleanly onto “takeaway” summaries, and the AI can misname entities or miss details. The transcript-to-knowledge pipeline also lacks some wishlist items: chapters don’t currently appear in the Readwise export, there’s no public API for developers, and a sustainable business model hasn’t been finalized (Snipd is free now, with “premium” being explored).

Even with those gaps, the workflow directly addresses three early frustrations with podcasts: slow consumption, limited skimming, and awkward note-taking. By turning audio into searchable text, then into snips that flow into Readwise and Obsidian, Snipd makes podcast learning feel closer to reading—fast to scan, easy to capture, and ready to revisit later.