Obsidian Plugins Qiaomu Reader and Qiaomu AI RSS Add TTS Voice Reading Support

On October 13, 2026, at 7:04 (shown in the post as 7:04 AM), author Xiangyang Qiaomu published a post on X saying that the Obsidian plugins Qiaomu Reader and Qiaomu AI RSS now both support voice TTS read-aloud. This is the only information source for this feature update; the post itself is the author's own account of a product update, not an official announcement or a third-party review.

The post also mentions that the day before, txt, html, and markdown rendering reading support was added to Reader. The author sums up the positioning of TTS as "if you are too lazy to read text, listening is a good choice," extending the reading scenario into the listening scenario.

It should be distinguished that the above feature descriptions are all the author's own statements and are at the source_claim level; the post does not include a version number, release timeline, or feature demonstration, so the status of "already supported" currently rests only on the author's unilateral statement.

According to the post, TTS read-aloud provides both cloud and local paths. On the cloud side, two models, Doubao Voice and Edge TTS, are added for reading articles and e-books; on the local side, a local small voice model is supported for real-time read-aloud.

The intent of this combination can reasonably be inferred as balancing quality and availability: cloud models usually depend on the network, while local models may fill in when the network is down. But the post does not explain the invocation method, quota, or cost of cloud TTS, nor does it explain the specific name, size, or operating requirements of the local small voice model.

In the comment section, user 0x小师妹 directly asks, "waiting for a specific name for the local model," and says that although running Edge cloud on an old machine at hand is stable, when the network is down they still want to use a local one to hold things over. This shows that the specific identity of the local model is an information gap readers care about, and the post does not provide an answer.

The author says both plugins have already been listed in the Obsidian official plugin library and gives download and installation addresses. But both addresses in the post are truncated to community.obsidian.md/plugins/qiaomu…, so the complete links cannot be confirmed, and the listing status cannot be independently verified from them.

Therefore, "already listed in the official plugin library" can currently only be treated as the author's own account. Whether the two plugins have indeed passed Obsidian's official review and been listed lacks first-party or third-party independent source corroboration.

The post's engagement data is: 7,276 views, 19 reposts, 1 quote, 22 likes, and 11 bookmarks. These numbers reflect the post's spread, not plugin installation counts or feature verification results.

The comment section provides two specific usage scenarios. User Pixel｜AI实用笔记 suggests importing a spoken script as text into Reader, choosing a section to listen to, and manually circling long sentences that cram three or four pieces of information into one breath and then splitting them shorter; this user also reminds that TTS reading smoothly does not mean a real person can speak it smoothly, and one still needs to read it through oneself to check pauses and stress.

Another user, Allen Chow, believes that turning an Obsidian knowledge base and AI RSS into listenable content allows continuous learning while commuting or running. Both usages come from commenters and are source_claim; the post author did not indicate adoption or official support.

The value of these discussions is that they suggest TTS's practical uses may go beyond "reading articles aloud," extending to writing checks and consumption of fragmented time; but they do not constitute technical verification of the feature itself.

Combining the post and comments, the information currently clearly missing includes: the specific languages TTS supports, the number of voices, and the offline availability scope; the invocation method, quota, and cost of cloud Doubao Voice and Edge TTS; the name, size, and operating requirements of the local small voice model; and the specific version number or release timeline of the newly added features.

In addition, whether Reader's newly added txt, html, and markdown rendering support was released simultaneously with the TTS feature is also not explained in the post. The author only mentions that rendering support was added "yesterday," and the sequence relative to the TTS launch cannot be confirmed.

These gaps mean that if readers want to assess usability in their own Obsidian environment, they still need to rely on the actual performance after installing the plugin, and cannot draw conclusions solely from the post's description.

From the verifiable information, this is a product feature update unilaterally released by the author: two Obsidian plugins add TTS read-aloud, covering cloud and local models, and are said to have been listed in the official plugin library. The post was published on October 13, 2026, with limited engagement, and no independent source has followed up.

For potential users, the key judgment point is not "whether TTS is supported," but the actual usability of the two paths, cloud and local: the local model name is unknown, cloud cost and quota are unknown, and the listing status rests only on the author's own account. Before these questions are clarified, it is more prudent to treat the update as "announced, pending verification."