WebDev

for row 266 about analytics dashboard, did you get the report verified by pankaj sir?

check the design there and send me you mail so i can trigger a mock mail to you

what about row 168 about chat session continuity? is it still on dev or has been moved to staging/prod

so the task is complete? can you verify it on prod

i am changing the status of task to pending till task updates column has been filled with: 1. how to enable it for other orgs 2. ⁠MR to staging 3. ⁠ss of how it looks like for 2 orgs other than elefant 4. ⁠ss of how it looks like for mobile view

made it for elefant we can turn it on for other orgs as well

its on “dev” server for all orgs?

its been resolved earlier also some of its fixes got merged in other MR's as sidefix

seprate one like daily update mails we sending weekly report mails not rolled out its in test sverer

row 215: "Send weekly ROI email to every client” has this become the new analytics page on dashboard? or is that a different task

@202611475136752 is this live on prod: “Fix conflicts in chat bubble rendering on dr malpani”: row 201. I am unable to verify because no MR links are there in task updates

no

currently on New embed design

Dashboard filters need to accept formula based filtering as well and option to save custom filtering with names

@111561876209776 do we need to rollout the v1 in staging?

Prompt management v2

@159738440466644 @30150821474450 @202611475136752 please do re-share your tasklist so I know what all tasks you've picked from the sprint

okay

worked on the daily email fixes for elefant, never tasked with enabling it for other orgs.

add this to the task updates of the relevant row

@159738440466644 please work on the MR for new embed design first; @202611475136752 @30150821474450 R lets not pickup any embed related tasks till new embed design is live. I will try to have it live today itself

@111561876209776 Email preferences toggles + tracking implemented and also tested locally. everything done and required is mentioned in file. please do check whenever you're available.

<media>

Is it already there or should we record it??

@30150821474450 for landing page blocker: instead of test links we can use "watch video demo”

Cool

<media>

Task well done! Good job guys

Please make this live as is.

Overall, 10/10. Kafi better lg rha hai

Pending Classification General Inquiry / New Chat — intent not yet known, excluded from the category breakdown below. 87↗ This is great. But there is no option to allow admins / sales people to re-classify the ticket to right category. Please add this in future scope. Basically ticket category edit krne ka option nhi dikh rha tickets page pe

Auto-resolved tickets kya hain?

2+1 = 3? Admin chat kya hai?

“Unattended (no AI or agent response)” Ye chats hyperlinked kyu nhi hain?

Same w “Chats Escalated to Agents”

In Elefant emailer: Chats Handled by AI (no human needed) -> This stat is wrong. So is AI Resolution Rate These should represent chats that were CLOSED / RESOLVED by AI successfully. Not those that got escalated to human agent. Abhi I feel agar agent ne join nhi kiya chat, to usko hum log resolved chat bol de rhe.

<media>

https://app.liveavatar.com/home

it can be done using a few videos itself

https://x.com/malpani/status/1899759228031742393

Here are two video demos I could find

they must have trained dr malpani’s avatar on liveavatar

https://x.com/malpani/status/1905442279277703238

Feedbacks / open points for video chatbot: 1. Can we use an Indian avatar or preferably Dr Malpani’s face? 2. The icons are a little confusing because there is a red mic with a cross and I'm not sure whether that means that I need to click on it before I start speaking or whether it's ready to listen 3. Is there any way the transcript can be shown while having the conversation? 4. Also there is quite a long lag between my pressing on the chatbot and her starting so I think you need to let people know that something is happening in the background

pusher is now setup on staging and dev

Suppression v3 done Suggested questions new architecture implemented Testing done for both above

@159738440466644 @30150821474450 EOD check-in Please reply with: 1. What did you complete today? 2. What's left / pending? 3. Any blockers? One message each is enough.

@111561876209776

no need to show the 3 qs again. “User clicked a suggestion question” is enough. The user’s message is already showing which option the user clicked

Flask is not being used in the video-call TTS path only via HeyGen

<media>

<media>

@111561876209776

https://beyondchats-voice-landing-page.netlify.app/

Mom : fixes in videochatbot , voice service landing page , prompt management status

Product Connect MOM 1. Answer Elefant Question on the sheet 2. Fix user suggested question architecture

It’s pronouncing correctly in Hindi just checked

https://dev.beyondchats.com/embed/demo/ivfindia.com/drmalpani.com

@202611475136752 for any AI message that is passed tyhru audio or video pipeline, we need to perform some amount of preprocessing on it because things like: “Rs" are not pronunciated properly. checkout preprocess_text_for_tts of background flask

https://meet.google.com/nqk-rtmw-gew?authuser=0

@159738440466644 joining?

in meet

right now, it should be on your google calendar

when?

@30150821474450 @159738440466644 @202611475136752 hi team, we have a product connect

we can take it up in the next sprint

okay ive made that a separate task then; row 435

no custom date need the schema changes currently dropdown since jan this year is included

@202611475136752 https://gitlab.com/Shoney13/beyond-chats-dashboard/-/merge_requests/235 does this also have custom date selection for analytics report?

okay thanks

Ticket-created timeline + live-chat AI-recovery-on-tab-close + minor UX. Zero overlap with the timeline-intelligence work.

is this different from timeline intelligence task?

Its not in git yet , i haven't created a repo since we are having a blocker regarding it

updates*

can you share access on gitlab to sid6i7 (me) and Pankaj-Baranwal (pankaj sir)

upload to notion and share the doc please

Also my brand guidlines is of beyondchats.com

landing page file in .md i added the instruction which is being followed by agent before creating or designig anything

In my local since landing page is a seperate new repo

where is it?

Yes

@202611475136752 @30150821474450 you guys have must created a brand guideline doc for beyond chats right? before working on landing pages

Agent joined, agent left, agent created ticket.

what all events are we covering?

backend: https://gitlab.com/Pankaj-Baranwal/chatgpt3_backend/-/merge_requests/890 dashboard: https://gitlab.com/Shoney13/beyond-chats-dashboard/-/merge_requests/245

yes please handle the greeting being spoken twice

Did you checked the other approaches on it ?

Let’s work on it if there’s no other priority task and see how better we can make it

and are you sure you tested it? Start karte hi the welcome msg is being spoken twice

I also putted the some changes yesterday night regarding the pause it didn’t taking before when user ask something in middle

actually even I had made some changes for optimizing video bot so I am not sure what the current state is

We having different accent bhi lip sync is better than the staging one and also no glitch I found

https://dev.beyondchats.com/embed/demo/ivfindia.com/drmalpani.com

And the test server changes did you check

didnt receive any feedback on it yet

@111561876209776 do we need any workaround on this ?

@159738440466644 updates?

Eod checkout : Landing Page Done, Working on videochatbot smoothness and perfect lip sync testing in test server

<media>

<media>

4bdd0a9 feat(video-call): wire text-mode lip-sync path in HeyGenAvatar

I dont see your changes in the dev branch

@111561876209776 check this here i switched to the text mode now the heygen voice is getting used and lip sync are better than before

https://dev.beyondchats.com/embed/demo/ivfindia.com/drmalpani.com

@111561876209776

https://jovial-pika-f7eb98.netlify.app/

great!

added reason

<media>

<media>

okay, please share the reasons first incase any updations are required

but we can add a reason mapped with intents

exisitng only intents is there

it will be helpful for cases like these. its not clear from the ss why it was suppressed.

do we store any “reason” field that we can show on the dashboard?

<media>

<media>

<media>

<media>

@202611475136752 updates?

Hi team heading out for some personal work will be back by 4

coordinated over call

we getting response from open ai model in both audio and text at the same time for the s3 recordings purpose we using the audio mode earlier we can try to switch it to text and test in test server regarding the smoothness first

yes we sending the audio output from the openai model to the heygen

and why did we choose voice model over text mode?

did you verify if the correct is correct?

but s3 audio may compromise after this change as we deal with the text and we need to record the heygen output may fail in ios

if this is priority can switch to this before landing page

Current Flow: User talks → OpenAI Realtime generates AI voice → we stream raw audio to HeyGen → HeyGen guesses lip movements from the audio → shows avatar video. Why lip-sync looks bad: We're using HeyGen the wrong way. Feeding it raw audio forces it to guess mouth shapes from sound waves — that's why mouth lags, breaks mid-sentence, and feels fake.HeyGen is built to take text, not audio. When you give it text, it does its own voice + perfectly matched lip-sync as one unit. That's what Exthalpy and every "real human" demo does. @111561876209776

@202611475136752 can you take a look at the architecture and answer this

@202611475136752

Why is the lip-sync so unnatural? I know for a fact that Heygen led real-time conversation is really good. Almost 1 year ago, Exthalpy - a competitor startup which is no longer in service had made a really good version live for Malpani clinic

And they had done it for Dr Malpani’s own avatar, lip sync was really good, voice was also Malpani cloned

Our version seems almost un-usable - it’s that bad.

I think meta webhooks are pointed to the prod server right? Will the dev server receive them?

Yes it's live on dev need some business accounts to test. Along with some small fixes as well.

messenger/instagram task has been pending for a long time. Are you facing trouble connecting accounts for testing?

Except for the Instagram OAuth issue, everything will be live today.

@159738440466644 updates for these?

tasklist: - timeline staging MR. - ⁠testing messenger/instagram task. - ⁠Instagram OAuth issue on dev.

Today’s task

@202611475136752 is landing page for video calls complete?

https://gitlab.com/Shoney13/beyond-chats-dashboard/-/merge_requests/235 Monthly report dropdown for the past months currently custom dates are not been there in this Mr cause that need the schema changes will sort in pahse 2

@111561876209776

User Suggested QS task

https://drive.google.com/file/d/12GGfegKiT36WLnQd0qJGkzCUCqiOCO3W/view?usp=sharing

every direct-count stat in Elefant's daily email clickable (16 stats — Chats + Tickets with filters pre-applied) , computed averages/% deferred to next phase.,https://gitlab.com/Pankaj-Baranwal/chatgpt3_backend/-/merge_requests/876 https://gitlab.com/Shoney13/beyond-chats-dashboard/-/merge_requests/232 @111561876209776

nothing from my personal account

arey i mean are they from your personal account or not

from dahsabord

@202611475136752 were the API keys for heygen and liveavatar provided by pankaj sir?

thanks

@111561876209776

rzp_live_rgP3z7WwP6Ydbk

From dashboard

did pankaj sir provide you with one?

@202611475136752 for razorpay, where can i get the REACT_APP_RAZORPAY_KEY=

Worked on landing page Ai suppression Blocker landing page not sure where to redirect for whatsapp, ig , messenger

Mr created mail audit and created mr for the dashboard toggle

Thanks will check it!

Need to test everything on the dev server regarding the messenger task. And do fixes if anything goes wrong.

@159738440466644 @30150821474450 @202611475136752 EOD check-in Please reply with: 1. What did you complete today? 2. What's left / pending? 3. Any blockers? One message each is enough.

checking

https://gitlab.com/Shoney13/beyond-chats-dashboard/-/merge_requests/226

@111561876209776 can you please help me with this config setup. this is happening on dev.

<media>

@111561876209776 daily email audits

Mr80 is old never merged fully covered by our recent ones 864,20,167 safe to close and toggle on chat config page isn't built yet only the video pill and banner are done on dashboard ,DB column is in 864will raise a small dashboard MR for the toggle after these 3 merge.

https://gitlab.com/Pankaj-Baranwal/chatgpt3_backend/-/merge_requests/869

what about this one?

All

@202611475136752 can you share all the MRs for video bot?

@102095734710467 what should be the test links for landing pages of features like instagram, whatsapp etc should we point them to social media accounts of beyond chats?

not sure

what abt others

<media>

<media>

and yes we will have a new sections for landing pages on our website that will point to these pages

all landing pages need to be static

@111561876209776 we are adding the landing page to beyondchats.com ryt? Can i get access to that repo ?

MOM - Task updates - ai suppression/no response task requirements

- pusher fixes. - ⁠email tracking phase 1 and 2

Work on the email audit Work on the email linked to dashboard filters in 2 phases for first one which is link able Add the months till jan so user can see the monthly report from jan month along with custom date range of its choice Heygen landing page

@30150821474450 @159738440466644 @202611475136752 can u share minutes?

https://beyond-chats.netlify.app/elefant-687e32d809004/analytics

Meet link or call?

Let’s connect after daily updates

@202611475136752 when can we connect?

@159738440466644 joining?

cool i will coordinate with nithin and pavan till then

Lunch 🥙

@30150821474450 @159738440466644 @202611475136752 Hi team, we have a product connect at 2. Can you confirm your availability?

fixed empty string audio urls were getting rejected by validator now coerced to null before validation and the first ai issue is not valid cause whereBetween is inside if (started_at && ended_at) guard so null can't reach it

there’s a blocker for that task (the event id thing you mentioned that day) that’s why I’ve not included those changes earlier.

@159738440466644 for the pusher fixes, I can see MRs for dev but not for staging: - https://gitlab.com/Shoney13/beyond-chats-dashboard/-/merge_requests/183 - ⁠https://gitlab.com/Pankaj-Baranwal/chatgpt3_backend/-/merge_requests/815

no?

Is there conflicts

Adding a LiveChat toggle in org config is a separate task

- check the 1st AI review in laravel MR, it should be taken care of: https://gitlab.com/Pankaj-Baranwal/chatgpt3_backend/-/merge_requests/864

Done: - ⁠Media messages implementation. - ⁠Track -> account owner first time replying to a lead. Pending: - origin implementation UI + API (campaign/device/AI) Blocker: - Do we need to have something like _LiveChat?_ toggle in org config, in order to let the account owners directly reply from their device?