Bugs Fixed:

1) Improved reliability for large file uploads:

Uploading larger files (like PDFs) to OpenAI Assistants was failing due to a 5-second default timeout, causing errors.

Hence we fixed it by increasing the receive timeout to 20 seconds and refactored the upload functionality, so that larger files can now be uploaded without errors.

2) Multiple URL buttons disabled in CTA section:

We identified an issue in HSM Messages where users could add only one URL button and one phone number button under Call to Actions. The option to add a second URL button appeared disabled, preventing multiple URL CTAs from being configured.

This restriction has now been resolved, allowing users to add more than one URL button in the Call to Actions section, as expected.

Enhancement Done:

1) Migrating Speech Processing from Bhashini to Gemini:

To improve reliability and performance, we have migrated Glific’s speech processing workflows from Bhashini to Google’s Gemini API (Gemini 2.5 Pro).

Over the past few months, Bhashini had been inconsistent, prompting this upgrade to ensure a more stable and scalable experience.

What’s Improved

• All Speech-to-Text (STT), Text-to-Speech (TTS), and translation workflows now run on Gemini
  
• Full support for:
  
  • Same source and target language
    
  • Indic ↔ English
    
  • Indic ↔ Indic language combinations
    
• Existing APIs continue to work with backward compatibility
  
• Enhanced metrics tracking for better visibility into usage
  

This migration ensures more reliable multilingual speech support across Glific while maintaining feature parity with Bhashini.

2) Standardising Icon Order on HSM and Triggers Pages:

To improve UI consistency and usability, we have standardised the order of action icons across the HSM and Triggers pages.

Previously, the View icon appeared first on the HSM page, while the Copy icon was first on the Triggers page, which could confuse users.
Now, the same icon appears first on both pages, providing a consistent interaction pattern and smoother navigation across Glific.

3) Extended Contact Profile History:

We have enhanced the Contact Profile History to give users better visibility into past interactions. Earlier, only the last 25 actions were visible. Now, the history is available for up to two months.

This extended history view helps users easily understand contact activity over time, making debugging and analysis much simpler.

Important Update:

We are excited to announce the Glific AI Chatbot Accelerator — a funded, six-month program for 30 selected NGOs to pilot AI-enabled use cases on their WhatsApp chatbots and turn innovative ideas into real-world impact.

To make it easier for you to join, we have rolled out a banner on your Glific instance login page.

If you would like to participate, you can register via the banner or contact us at tanu@projecttech4dev.org.Don’t miss this opportunity to explore how AI can enhance your chatbot workflows and drive meaningful results!

Documentation Update:

Revamped Documentation:

1)  Media & Attachment Guidelines:

We noticed that many NGOs were facing confusion around which media types and sizes can be uploaded in the Send the contact a message node.
To make this clearer, we have updated our documentation with all the details.
Do read the updated documentation here


2) Facebook Verification Video Tutorial:

To make onboarding easier for new NGOs, we have added a dedicated Video Tutorial section for Facebook verification.

Many NGOs requested a more detailed, step-by-step video format to guide them through the process, so this update provides a clear and easy-to-follow reference to complete verification successfully.

Check out this video for full guidance.