Short answer
RaftLabs built Voter IQ, a Clubhouse-style live audio app for political networks in Andhra Pradesh, India. The first iOS, Android, and web admin release shipped in 16 weeks, ahead of an election cycle. In testing, the app handled about 7,854 people connected at the same time, and answered 1,000 simultaneous users in about a third of a second.
Engagement
The Voter IQ engagement
- Client
- Voter IQ, a founder-led civic tech startup in Andhra Pradesh, India
- Decision-maker
- The founder, a product-minded operator with deep knowledge of how the region's political structure works
- What it is
- A live audio app where party leaders host discussions, badge holders follow their leaders, and citizens listen, comment, and vote in polls
- What we owned
- iOS and Android apps, the web admin panel, the behind-the-scenes system, and crowd testing
- Why RaftLabs
- Our real-time voice work for PSi, a group deliberation platform
- Team
- One backend engineer, one mobile engineer, one QA engineer, one designer, and a project manager
- Engagement model
- Phased product build
- Duration
- First release in 16 weeks, followed by further product phases
The situation
In Andhra Pradesh, political conversation follows a ladder. MPs and MLAs sit at the top, then constituency leaders, then badge-holding party members, and finally the citizens they all represent. Voter IQ's founder wanted every rung of that ladder in one app. Leaders host live audio rooms, members follow the leaders they answer to, and citizens tune in to listen, comment, and vote in polls.
A WhatsApp group can't do that. It has no idea who the local MLA is, which constituency a member belongs to, or who should hear about a broken road.
And that matters because most voters meet their down-ballot candidates for the first time on the ballot itself. A network that knows your constituency is the difference between participating and guessing.
The timing mattered too. Elections were coming, and the founder wanted the app in people's hands before they did.
The founder came to us after seeing the live voice platform we built for PSi. Sixteen weeks later, the iOS app, the Android app, and the web admin panel were live. Before calling it ready for crowds, we put it through large simulated audiences to see how it held up.
starting point
From scattered groups to one structured network
Build an app that understands the political hierarchy, puts each person in the right place within it, and is ready before polling season.
- Political discussion moved through scattered groups and calls, one level at a time
- Getting the right leaders into a conversation meant manual outreach and reminders
- Citizens followed politics through general public channels, not their own constituency
- Generic tools had no idea who was a host, speaker, badge holder, citizen, or admin
- Local issues had no clear route to the constituency leader, MLA, or MP
- Live audio rooms with host, co-host, speaker, and listener roles
- Four types of users, each with their own permissions on mobile and web
- Citizens listen, comment, and vote without leader-level access
- Groups tied to each constituency, with feeds that show the sessions and issues relevant to each person
- Issues move up to the next leader automatically if they aren't picked up
- Tested before launch with thousands of people connected at once
The problems worth reporting
Every feature depended on the hierarchy
Who can see a room, which issues show up in someone's feed, where a complaint goes next: all of it depends on where a person sits in the party structure. So before anyone wrote code, our PM sat down with the founder for several sessions to map how parties, constituencies, and leadership levels connect.
The founder knows this world from the inside. As early versions came together, the founder and our team kept refining the structure together until the app matched how these networks work on the ground.
Constituency reshaping was deferred so launch wouldn't slip
Members, leaders, groups, issues: almost everything in Voter IQ is attached to a constituency. Reshaping that structure, for example removing a constituency, ripples through all of it, which made it the biggest schedule risk in the build.
We gave that work its own phase after launch. The founder agreed, and the first release went out on time.
Turning a political hierarchy into app rules
Every user had to land in the right party and constituency, get the right powers, and see a feed of sessions and issues that fit their role and area.
We kept four things apart: which network someone belongs to, who they follow, what they can do in a live room, and whether they can manage other people. That way, following an MLA doesn't quietly give you an MLA's controls. Party admins, badge holders, citizens, and web admins each see exactly what their role allows.
Issues that climb the ladder on their own
When a constituency leader raises an issue, it goes to a named leader one level up, with a deadline attached. If that deadline passes without action, the app sends it up another level.
That gives every local problem a clear route toward the MLA and MP, and keeps it moving until someone picks it up.
Live audio ready for a crowd
A political discussion can fill up fast, especially close to an election. We used Agora, an established live audio service, to carry the sound, and spent our effort on the rules for who can host, speak, listen, comment, and vote.
Before launch, we simulated large crowds hitting the app at once. With 1,000 people active together, every action went through and responses came back in about a third of a second. The live connection layer handled about 7,854 people at the same time on one server, and more servers can be added as the audience grows.
Proof
Results
Delivery and testing results from the Voter IQ build.
| Result | What changed | Period or context | Evidence and limitation |
|---|---|---|---|
| First release | 16 weeks to iOS, Android, and web admin | Live before the elections | Published on Google Play and the App Store |
| 1,000 people at once | Every action went through, in about a third of a second | Simulated crowd test before launch | RaftLabs test report |
| Live connections | About 7,854 people connected at the same time | Tested on one server, with room to add more | RaftLabs test report |
the build
What we built
Live audio rooms with political roles built in
Leaders open a room, invite co-hosts, and bring people up to speak. Citizens listen in, comment, and vote in live polls. Agora handles the audio, and the app decides who gets to do what based on each person's place in the network.
One build for iPhone and Android
With the election date set, building two separate apps would have doubled the work. React Native let the team write it once and release it on both platforms. You can find it on Google Play and the App Store.
Constituency groups and issue escalation
New users join the group for their constituency, and their feed shows sessions and issues for their area and role. Issues from constituency leaders go to a named senior leader and move further up if nobody acts.
A web admin panel for running the network
Admins manage networks, constituencies, members, and permissions from a browser, so the team behind the app can run things day to day without a developer.
The lesson
Plan for a full room from day one
A live audio app gets judged at its busiest moment, when a popular leader goes live and everyone joins at once. Pick tools that can grow with the audience, and test with a crowd before real people arrive. On Voter IQ, leaning on Agora for the audio freed the team to focus on roles and permissions, and the crowd tests gave the founder real numbers to plan around.
How it runs
- React Native
- With an election date fixed, there was no time to build two separate apps. One shared build shipped on iOS and Android in 16 weeks.
- Agora
- Carried the live audio, so the team could focus on political roles and permissions. The app still decides who can host, speak, and listen.
- AWS
- Runs the app behind the scenes, and can add more servers as the audience grows.
- Firebase
- Handled push notifications for upcoming sessions and tracked how people used the mobile app.
Common questions
RaftLabs built Voter IQ, a Clubhouse-style live audio app for political networks in Andhra Pradesh, India. Leaders host rooms, badge holders follow their leaders, and citizens listen, comment, and vote in polls. The app also has constituency groups, issues that move up to senior leaders automatically, and a web admin panel. The first iOS, Android, and admin release took 16 weeks.
Voter IQ's first release took 16 weeks with a small team: one backend engineer, one mobile engineer, a QA engineer, a designer, and a PM. If a service like Agora carries the audio, the live rooms come together quickly. Most of the time goes into roles, permissions, and admin tools. Our guide to building a social audio app like Clubhouse covers the rest.
That depends on how it's built, which is why we test before launch. In Voter IQ's tests, 1,000 people used the app at once with responses in about a third of a second, and one server held about 7,854 live connections. Adding servers raises that ceiling as the audience grows.
Treat membership, following, speaking rights, and admin powers as separate things, so following a leader never hands someone that leader's powers. Voter IQ has four types of users, each with its own actions, all organised around constituencies. We mapped the structure with the founder over several sessions before building, then shipped it on iPhone and Android from a single React Native build.
A constituency leader raises an issue, and it goes to a specific senior leader with a deadline. If the deadline passes without action, the app moves the issue up another level, toward the MLA or MP.
Yes. If you're building a live audio, community, or member-network product, take a look at our social media app development and real-time application development services.
Related work
More real-time voice work
300+ people in one anonymous voice discussion
PSi runs real-time voice deliberation that splits large groups into small tables. Voter IQ's founder hired us because of this build.
3,000+ audio and video minutes in two beta weeks
Worxwide's hybrid work app put meetings and virtual office rooms in one product for distributed teams in India.
3,500 field staff completing training every day
Eris Lifesciences' EMS Connect reached 3,500+ daily users across India, with offline access for field reps.