From finding an astrologer to a paid consultation
Native Android and iOS development in Kotlin and Swift for Nakshatra, a product connecting customers with astrologers.
A consultation marketplace needs to coordinate more than messages: who is available, whether a customer can fund a session, when the expert accepts it and how the session ends. Nakshatra brings these decisions into a mobile experience with separate customer and astrologer roles, supported by an administration panel.
Screens are selected from the supplied 25-page interface-design PDF, not captures of a running app. Names, balances, ratings, prices and dates are example content. Implementation details below come from the historical Android handover; they do not independently establish iOS feature parity. Kotlin and Swift are owner-confirmed; the older handover includes Java snippets. The live-session design is broader than the documented YouTube-based implementation.
Original interface designs, with implementation context from the project handover.
Phone, Google and Facebook authentication provide the account entry points. New customers complete a profile with date, time and place of birth; returning customers continue to the home screen. Customers can later edit their own details.
The wallet groups the available balance, recharge entry and transaction histories in one place. Razorpay is named in the Android implementation; the PDF's payment-method labels are design references, not evidence of the final provider configuration.
The chat view separates the message timeline from the composer and delivery status. The original designs also show a conversation list with message previews and unread counts, giving returning customers a route back to their exchanges.
Separate call and chat histories organise previous consultations. The chat-history design pairs the order reference and date with duration, per-minute rate and the amount deducted, plus controls to view the conversation or start another chat.
Reviews are public on the astrologer's profile, but the rating prompt appears only after a chat ends. Customers can submit a rating or skip the prompt, keeping feedback tied to the consultation lifecycle.
The Android handover describes upcoming, recent and live schedules that open sessions on YouTube. The original PDF explores an in-app viewer with comments and gifting; those richer designs are not presented here as delivered streaming or payment functionality.
The Android handover documents a wallet check and an explicit request-and-accept sequence between the two apps.
The customer must hold at least six times the astrologer's per-minute chat rate. An insufficient balance opens a recharge prompt. This is the recorded start condition, not a claim of a six-minute minimum charge.
Firebase Realtime Database records the requested state and a notification reaches the astrologer. A rejected request returns the customer to the home screen with a message.
On acceptance, the astrologer app creates a chatroom through the server API and in Firestore. The customer's listener observes the accepted state, joins the room and starts the timer and heartbeat.
Separate states identify whether the customer or astrologer ended the chat. The end-session flow stops the heartbeat and moves the customer toward feedback.
Closing an app or losing connectivity can bypass its normal end-chat action. The implementation therefore gives the server a recurring signal that the session is still active.
The handover describes a roughly ten-second ping cycle. Its sample schedules the next request nine seconds after a successful response and passes a ten-second waiting value to the API.
If those pings stop arriving, the documented server behaviour terminates the chat and deducts the session amount. A server rejection or client request error also enters the end-chat procedure.
Realtime Database carries request and session-state changes, Firestore participates in the conversation record, and Cloud Messaging delivers device notifications. REST endpoints coordinate chatroom creation and heartbeat handling.
Expert profiles bring languages, experience, skills and consultation rates together. Astrologers have Online, Offline and Busy states; an active chat changes their status to Busy. Their profiles are edited through the administration panel. The documented call action invokes a calling API with the two parties and a maximum-duration parameter, and saves call data in Firestore.
The 2020 Android report flags unresolved server-related chat disconnections and marks Panchang as disabled at that time. A later fix, current release status and measured business outcomes are not established by the supplied material. This case study describes the documented product and its engineering decisions at that stage.
Bring us your idea. We’ll shape the next step together.
What our clients value about working with Daphnis Labs.
The team at Daphnis Labs redefined what’s possible for Urbanface. They delivered a bespoke, animation-heavy website that remains incredibly quick and functional. The…
We wanted a unique, 'one-of-a-kind' feel for Glareen, and Daphnis Labs delivered an ecosystem that is both beautiful and technically superior. Their expertise in…
The level of technical depth Daphnis Labs brought to our Game project is unparalleled. While the front-end reel games are visually stunning and highly engaging, the…
Practical perspectives on AI, product engineering, commerce and modern software delivery.
Measure recovery by restoring into an isolated environment and checking the application, roles and dependencies that need the data.
Read Blog
Classify responses before caching them, make cache keys reflect their audience and test what happens when permissions change.
Read Blog
Review database roles, background jobs and exports together when designing row-level security for a shared application database.
Read Blog
Plan rollout, fallback behaviour and flag removal together so temporary release controls do not become permanent product complexity.
Read Blog
Connect traces, metrics and structured logs around a real failure path so the team can locate an incident and choose a next action.
Read Blog
Define permissions around concrete actions, and make an approval apply to the exact message or record that will be changed.
Read Blog
Separate parsing, validation and approval so an ordinary spreadsheet upload does not become an opaque bulk edit.
Read Blog
Give images, third-party scripts and interactions measurable limits, then investigate regressions by page template and device.
Read Blog
Help people complete a form with persistent labels, specific errors, preserved answers and a clear confirmation state.
Read Blog
Separate event receipt from order processing, record duplicate deliveries and recover work that stops halfway through.
Read Blog