In this build, language switching has been optimized and the language conflict bug has been resolved. Along with that, audio has been further optimized and improved, and UI has also been improved.

We have also added Rule functionality, through which you can create your own custom rules. This is a very powerful feature that allows you to add your custom rules to the app. For example, I am sharing a rule below so you can refer to it and create your own custom rules.

This is a rule file that I am sharing with you all as an example. You can check it out and create your own custom rules in the same way — for emoji processing, text processing, or for reading any text with any engine you want.

You can also create custom rules to manipulate strings in any way you want — for time, date and time, number processing, text processing, to detect any pattern and make it speak in any way you want.

This is an example of camelCase processing. You need to import it by going into the Rules settings, activate it, and also turn on / enable its master switch.

Optimize the switching audio sample rate. Make it more optimized to get the maximum quality of the switching voice on the rules menu you can read the full documentation on how to create the rules and export the prompt and document

In this build, we have fixed a lot of things — especially the audio mismatch / sampling rate issue. Along with that, we also fixed the words ignore issue, fixed that bug, and fixed the optimization issue that was occurring on some devices.

I especially want to thank everyone who is reporting issues and identifying each issue like an expert — reporting the bugs so well, just like specialist testing engineers.

While many of you are sending me messages, suggestions, and bug reports, I try my best to read all of them. However, it is not possible for me to reply to everyone individually. Please be patient — I am not ignoring your messages and suggestions.

Thank you all very much. Please continue to support us in the same way and keep reporting the bugs. With your help, we will make it even better and keep improving it continuously.

Improve the language switch logic and optimize overall performance. Also add the date reading format. You can check under the text processing.

The latest Jieshuo+ update has been pushed directly via OTA.
Paid users, please check your notifications to download the latest version. Also, don't forget to check the official changelog.
New users who wish to purchase the Jieshuo can contact us.

What’s New in IBM Eloquence: The Speed & Clarity Update

We’ve completely revolutionized the delivery pipeline powering your favorite voice. While the classic Eloquence sound remains 100% authentic and untouched, the underlying execution has been re-engineered for absolute maximum performance on modern Android devices.

In rigorous physical hardware testing, IBM Eloquence beats official Code Factory Eloquence benchmarks by an exceptional margin.

Here are the certified performance metrics behind your new screen reader experience:

🚀 Over 90% Faster Touch-to-Speech Response
Standard pipelines trap audio behind deep system queues, introducing noticeable hesitation. We have completely eliminated this delay. Audio hits your device's speaker in just 4.0 milliseconds—delivering an instantaneous response that outperforms official Code Factory Eloquence by an exceptional margin.

⚡ 22% Faster High-Frequency Navigation
Power users demand speed. In continuous stress testing under high-frequency navigation demands, IBM Eloquence processed interruptions 22% faster than official Code Factory Eloquence. With an ultra-responsive turnaround time of just 0.67 milliseconds to drop an utterance and speak the next, you can glide through complex menus without a single stutter or hesitation.

🎧 Flawless, Zero-Jitter Audio Delivery
Say goodbye to random audio cracks, micro-stutters, and clicks. Our upgraded architecture ensures a continuous, crystal-clear audio stream, running completely free of dropouts even during peak background workloads.

🔋 Primed Hardware Wake-Up (No Clipped Syllables)
Legacy setups let phone speakers enter a deep sleep between actions, causing the first syllable of the next item to be clipped or muted. Our intelligent idle management keeps the audio hardware primed and ready during active navigation, guaranteeing crisp, complete words from the very first letter.

🛡️ Bulletproof System Stability
Your screen reader is your lifeline. We have engineered intelligent, real-time fail-safe mechanisms directly into the engine. In the rare event of an unexpected processing hiccup, the system recovers instantaneously, ensuring you are never stranded in silence.

Don't just take our word for it—test it yourself.
We believe benchmark numbers should speak through everyday use. Update the app, turn on your screen reader, and navigate through your heaviest interfaces at top speed. Experience firsthand how IBM Eloquence outperforms official Code Factory Eloquence to set a new standard in responsive navigation.

Stay tuned for ongoing improvements. Keep reporting any quirks or edge cases you encounter so we can hunt them down and crush them immediately—the more testers pushing the limits, the more rock-solid and unbreakable this engine becomes for everyone.