GirlScript Summer of Code 2025

[eccentriccoder01]

Welcome to the Official Discussion Space for TalkHeal - GSSoC'25 Folks!

Since the GSSoC team hasn’t yet set up project-specific Discord channels (understandable with the scale of contributors), this GitHub Discussion thread will serve as a dedicated space for all TalkHeal-related communication until then.

---

What You Can Use This Thread For:

• 🤝 General questions or doubts about the project
• 🤝 Suggestions for new features
• 🤝 Help with open issues, PR guidance, or debugging
• 🤝 Coordination between multiple contributors
• 🤝 Sharing learning resources or ideas related to TalkHeal (Streamlit, Python, NLP, UI, etc.)
• 🤝 Updates on feature progress (especially large/Level 3 issues)
• 🤝 Anything else relevant to collaboration on TalkHeal and other Open Source related info

---

💙 About TalkHeal:

TalkHeal is an empathetic, intelligent, and interactive mental health support assistant built using Python and Streamlit. Designed with compassion and care at its core, it offers 24/7 support, emotional journaling, resource guidance, and AI-powered conversations powered by Google’s Gemini

Feel free to work on existing unassigned issues or create new ones, and check out my other GSSoC25 projects too:

• 
• 
• 
• 

🔧 Projects I've Submitted for OSCI

We can collaboratively participate in OSCI and other upcoming events. As a project admin for OSCI, I have submitted 2 intermediate level projects... Check these out as well so it might help you contribute from before and get to the top of the Leaderboard:

• 
• 

Stay tuned for upcoming Open Source events too!

📩 Let’s connect:

🌱 Join the EduLinkUp Developer Community

We’re building an ecosystem of student developers through EduLinkUp. Join our developer community in the EduLinkUp ecosystem! We discuss open-source events, share resources, get help, and collaborate. Let's grow together. 🚀

---

Feel free to introduce yourself below!
Looking forward to building something truly helpful and meaningful together 💜

---

[eccentriccoder01]

Authentication Work

Several contributors are currently working (or have expressed interest) in improving the authentication flow:

Active Work:

• @PragnyaKhandelwal – Issue Fix: Hide Sidebar and Chat UI on Login Screen #71
• @jyothika-badugu – Issue Feature Request: Add Login Page to the Website #55 (originally opened by @sirapurapushyam)

And couple of others too who have opened issues but then backed out of working on it.

🕓 Inactive/Pending Issues:

• Issue Improve Authentication, Chat UI Stability, and User Experience #16 by @MaddulaSamanvitha
• Issue Feature Request: Add Mood data Visualization via graphs and charts #11 by @Baibhab-Adhikari

🔐 The current login functionality was implemented by @srishti1837 in PR #61, and she has confirmed that it currently restricts access only to the Journaling section, while the chatbot and other areas remain accessible when logged out, as the feature was scoped specifically to journaling.

Contributors working on Authentication: are welcome to explore and choose any of the following approaches:

• Place the login UI directly inside the Journaling section for clarity and a focused experience.
• Extend/enhance authentication to cover the entire app.

There’s no strict rule, you can create new issues depending on the scope you're working on. Just make sure to discuss and decide among yourselves to align and avoid duplicate efforts or conflicts. I’ll be happy to assign issues to you accordingly.

Also I'll try to keep all authentication based issues/PRs at level 3.
Let me know if anything’s unclear or if you'd like feedback on a particular direction.
Thanks for contributing to TalkHeal!

[parth-patel010]

Hello @eccentriccoder01

[eccentriccoder01]

Yes @parth-patel010

[AvichalTrivedi7]

Hiii, Avichal here
I'm really passionate about AI and helping each other in order for us all to grow together :)
I really like how well we resonate with each other and i wanted to contribute to this meaningful project.
It would be lovely if you could suggest me something that i could work on, I have worked on LLMs alot and would love to know more about them.

PS: sorry to write so much and ask you about how i can help, but i was really lost even after reading all about the current problems and not knowing where to start...
If you could guide me a little, it would be wonderful

Thankyou so much !

[eccentriccoder01]

Great to see your passion for AI and LLMs Avichal

Since you’ve already got experience working with LLMs, a good place to start would be around improving the chatbot’s conversational depth and personalisation. For example, you could look into:

• 🔹 Long-term memory integration — using embeddings & vector databases so the chatbot recalls past interactions.
• 🔹 Persona-aware responses — making the chatbot adapt its style to the user over time.
• 🔹 Evaluation & fine-tuning — testing conversation quality and suggesting improvements.

Explore the repo, take a look at the open or unassigned/unprogressed issues in the ISSUES section or create and let me know what sparks your interest... I’d be happy to guide you further and assign you something concrete 👍🏻

---

[AvichalTrivedi7]

Will do,
Thankyou so much

[AvichalTrivedi7]

@eccentriccoder01
Just sending a follow up with a mention in case my message got accidentally overlooked.

[Shrutiso]

Hey @eccentriccoder01
I’m Shruti Somvanshi, and I’d love to contribute to your project Saral Seva. I was really inspired by the concept, and I’d like to add an extra edge to it by integrating a new feature around understanding standard legal terms.

This idea is also connected to my own project, Kannon Sarthi, which focuses on simplifying legal understanding for people. I believe combining this with Saral Seva can make the platform even more impactful and user-friendly.

[TruptiGhanvat]

Hi Avichal, Sorry for the late reply — I was busy with exams, so I saw your message a bit late. It’s really nice to see your passion for AI and LLMs.Don’t worry about writing long — it actually shows your genuine interest and clarity. For contributing, you can start with: Exploring fine-tuning / alignment techniques for LLMs to make them more domain-specific. Working on evaluation benchmarks for fairness, bias, and efficiency. Or picking up a small module in our project (like preprocessing, chatbot interface, or knowledge integration) where your LLM expertise can be really valuable. If you’d like, we can set up a small discussion to align your interest with the project needs so you can get started smoothly You can also reach me directly at 9689321157. Looking forward to collaborating with you!

…

On Fri, Sep 26, 2025 at 12:58 AM Avichal Trivedi ***@***.***> wrote: Hiii, Avichal here I'm really passionate about AI and helping each other in order for us all to grow together :) I really like how well we resonate with each other and i wanted to contribute to this meaningful project. It would be lovely if you could suggest me something that i could work on, I have worked on LLMs alot and would love to know more about them. PS: sorry to write so much and ask you about how i can help, but i was really lost even after reading all about the current problems and not knowing where to start... If you could guide me a little, it would be wonderful Thankyou so much ! — Reply to this email directly, view it on GitHub <#68 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BTIIOMS3IASSD2XJGNNL6OT3UQ66RAVCNFSM6AAAAACCVYJKFKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINJRGQ4DIOA> . You are receiving this because you are subscribed to this thread.Message ID: <eccentriccoder01/TalkHeal/repo-discussions/68/comments/14514848@ github.com>

[AvichalTrivedi7]

Alright, will do.
Thankyou so much.

[TruptiGhanvat]

Hi Shruti, Sorry for the late reply — I was busy with exams, so I saw your message a bit late. Thanks a lot for reaching out and showing interest in Saral Seva 🙌. Your idea of integrating a feature to simplify standard legal terms sounds very impactful. I really like how it also connects with your project Kannon Sarthi. Combining both approaches could make Saral Seva not just a service platform but also a knowledge bridge for users. Let’s explore how we can integrate your idea — maybe as a legal assistance module where users can get simplified explanations of legal terms directly. You can also reach me directly at 9689321157. Excited to brainstorm this with you and make Saral Seva more user-friendly together!

…

On Sun, Sep 28, 2025 at 5:06 PM Shruti Somvanshi ***@***.***> wrote: Hey @eccentriccoder01 <https://github.com/eccentriccoder01> I’m Shruti Somvanshi, and I’d love to contribute to your project Saral Seva. I was really inspired by the concept, and I’d like to add an extra edge to it by integrating a new feature around understanding standard legal terms. This idea is also connected to my own project, Kannon Sarthi, which focuses on simplifying legal understanding for people. I believe combining this with Saral Seva can make the platform even more impactful and user-friendly. — Reply to this email directly, view it on GitHub <#68 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BTIIOMTGBSFY3RLEKRKHGRL3U7B5LAVCNFSM6AAAAACCVYJKFKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINJTGM3TAMQ> . You are receiving this because you are subscribed to this thread.Message ID: <eccentriccoder01/TalkHeal/repo-discussions/68/comments/14533702@ github.com>

[student-singhkamini]

@eccentriccoder01 hello sir I am Kamini, really interested in contributing TalkHeal ,My final year project similar to this project so want to contribute in this project.

[pkbros]

Welcome @student-singhkamini to Talk-Heal.
I am Prateek, mentor for TalkHeal during gssoc25.
Feel free to work on existing unassigned issues or create a new one.
Its great your project is similar to this one, your knowledge and experience will sure upgrade TalkHeal.
Good Luck!

[eccentriccoder01]

@eccentriccoder01 hello sir I am Kamini, really interested in contributing TalkHeal ,My final year project similar to this project so want to contribute in this project.

Sure Kamini, welcome to TalkHeal!

[student-singhkamini]

hello sir, I want to know one thing from where you take data for train the chatbot and in which part of the project I will have to work ?

…

On Tue, Sep 30, 2025 at 4:11 PM Prateek Sharma ***@***.***> wrote: You're welcome @student-singhkamini <https://github.com/student-singhkamini> to Talk-Heal. I am Prateek, mentor for TalkHeal during gssoc25. Feel free to work on existing unassigned issues or create a new one. Its great your project is similar to this one, your knowledge and experience will sure upgrade TalkHeal. Good Luck! — Reply to this email directly, view it on GitHub <#68 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BCFCEDCWKIXFU3MJBEVQ2YL3VJM4RAVCNFSM6AAAAACCVYJKFKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINJVGE2TCNQ> . You are receiving this because you were mentioned.Message ID: ***@***.*** com>

[LavanyaMurudkar]

Hi, Lavanya here!
I went through the live demo of the TalkHeal project and found it really impactful. I believe I could contribute meaningfully to the UI/UX design aspect, as I noticed areas where the user experience and interface could be improved for better accessibility and emotional connection.

I’m a final-year engineering student specializing in AI and Data Science, with hands-on experience in UI/UX design using Figma. I’d love to work on enhancing the overall design flow, visual consistency, and usability of the platform to make it more comforting and intuitive for users.

Could you please guide me on how I can start contributing to the design side of this project?

[Tejaswimadastu]

Hi, I’m Tejaswi 👋
I’d love to contribute to the project by enhancing the chatbot’s interaction quality.
One way to do this is by implementing persona-aware responses, where the chatbot adapts its tone and style based on the user’s preferences.
📝 Proposed Feature
Allow users to select or define a preferred persona style (e.g., formal, informal, supportive, neutral).
Integrate persona settings into the prompt/response generation layer.
Modify chatbot responses dynamically based on the selected persona.
✨ Why This is Useful
Makes the chatbot feel more personal and human-like.
Improves engagement and emotional connection with users.
Creates a better mental health support experience.
🧰 Implementation Ideas
Persona selection in the UI (dropdown or buttons).
Prompt engineering to reflect persona style.
Optional: persona detection based on user inputs.