Settings → Wallet now offers month-level CSV and PDF statement downloads — opening + closing balance, full transaction history, GST breakup when applicable.
Each statement covers one calendar month (IST), shows opening + closing balance, and lists every transaction with date, type, description, signed amount, balance after, and the Razorpay payment ID for topups. The PDF mirrors the same data in branded format suitable for forwarding to your accountant; the CSV imports cleanly into Excel or Google Sheets.
These are wallet statements, not GST tax invoices — if your business needs invoices for input-tax-credit claims, email ravi@citare.ai. GST-compliant tax invoices ship as a separate feature.
Every data table that's processed for you now offers a CSV download — AI Responses, Recommendations, Competitors, Monitored Queries, Sources, Services, Topics, Accuracy.
Each table header gains a small CSV button. Click it and you get a spreadsheet-ready CSV of that surface's data, scoped to your client account, with a header block summarising the period and any filters you'd applied. Files import cleanly into Excel, Google Sheets, or any data-analysis tool.
Daily limit is 20 exports per client across all surfaces (combined with the existing audit / KG / monthly report / wallet exports). The standard going forward: every dashboard table that we process for you must offer a way to export its data.
Settings → Brand voice now configures the tone Opus writes in for each client. Sample paragraph, tone descriptors, preferred / avoided terms, formality, pronoun convention.
Today every monthly report sounds like Citare wrote it. With brand voice, each client's report sounds like that client (or their agency) wrote it. Especially valuable for agencies managing multiple clients — the same dashboard generates AcmeCorp's report in AcmeCorp's house style and BrandLoom's portfolio in BrandLoom's. White-label credibility without manual editing.
The single most powerful field is the sample paragraph: paste 100–300 words of the brand's existing copy and Opus will match its rhythm and vocabulary. Tone descriptors and term lists are belt-and-braces.
Off by default. The feature is fully opt-in via a checkbox in the same panel; until that toggle is ticked the narrative prompt is byte-for-byte identical to before, no extra tokens, no extra cost. Even when on, the per-narrative cost increment is fractions of a paisa.
The Friday email digest gained a Directory Health section highlighting clients with unresolved issues from the bi-weekly directory check.
Aggregate count of clean / issues / unlinked directories shows in the email body, with a per-client breakdown for any client carrying unresolved issues. Clients with no issues are omitted from the breakdown to keep the email tight.
Monitoring runs now finalize as one of four states — completed, partial, insufficient_data, or failed — and below-threshold runs skip wallet charges entirely.
If a run fails to hit the platform-completion threshold (default 50% of expected samples per platform), Citare skips wallet deduction, skips recommendation generation, and flags the run for ops review. You don't pay for runs that didn't produce real data.
Partial runs that meet the threshold still finalize normally, with the partial flag visible across the dashboard so you know what you're looking at.
Set a balance threshold and we'll automatically top up your Citare wallet via your saved card before the next monitoring cycle.
Configurable threshold (default ₹1,600 — one weekly cycle), monthly cap to prevent runaway charges, and self-disable after 3 consecutive failed attempts. Card-based today; UPI Autopay coming when there's customer demand.
Disabled by default — turn it on under Settings → Wallet whenever you're ready.
Every AI response is now classified for brand sentiment + category sentiment. Dashboard chips show the polarity at a glance; monthly reports flag divergence between brand and category tone.
When AI talks about your business, the tone matters as much as the mention itself. Each response on the AI Responses table now shows a sentiment chip (Positive / Neutral / Mixed / Negative / Absent), and the dashboard Overview adds a Brand vs Category sentiment card showing both stacked side-by-side.
When brand sentiment diverges from category sentiment by more than 0.3 polarity points, a callout chip appears on the dashboard and a corresponding recommendation lands in the queue.
New diagnostic tab showing how each AI platform actually treats your queries — citation density, AI Overview suppression rate, generic-answer ratio.
ChatGPT, Claude, Gemini, and Google AI Overviews behave very differently on the same query. Some respond substantively, some default to generic advice, some don't surface AI answers at all. The new Platform Behavior tab on the dashboard surfaces those patterns per run.
It also feeds a new section in your monthly report — "How AI platforms behaved this period" — that explains the data in plain language without prescribing actions. The signal is the diagnostic; the recommendations remain in the existing recommendations queue.
The homepage now leads with what the dashboard actually shows you, instead of category education.
Above-the-fold is a real screenshot of the visibility score over time, the AI Responses table, and the recommendations queue. Below the fold goes deeper into individual capabilities. The redesign was driven by a sales feedback loop — agencies want to see the product before reading about the category.
Settings → Generated Presence now shows the auto-built business profile we feed AI crawlers. View, copy the public URL, regenerate at will.
Each Citare client gets a public profile page at /presence/<slug> structured for AI consumption — schema-rich, llms.txt-aligned, with About / Services / FAQ sections derived from your knowledge graph. Crawlers and AI assistants treat it as a clean signal source for grounding answers about your business.
Click any column header on the AI Responses table to sort — by date, platform, mention status, sentiment, position.
Sorting runs server-side so it works correctly across the full result set, not just what's loaded on the page. Sort state is preserved in the URL so you can share links with sort applied.
Four new marketing pages with role-specific value props: For Agencies, For Indian SMBs, For SaaS founders, For Content Teams.
Each page reframes Citare's capabilities through the lens of one ICP — what an agency principal cares about (white-label margin), what an SMB owner cares about (does AI know I exist?), what a SaaS founder cares about (am I cited on category-discovery queries?), what a content team cares about (which content actually gets cited?).
Linked from the homepage and from the Solutions footer column.
Agencies can now onboard and switch between multiple client accounts from a single login.
New /dashboard/clients/new flow handles agency-side client creation, and the dashboard top bar lets agency admins switch between clients without logging out. Client-account isolation is preserved — agency admins see only their own clients, never another agency's data.
Want updates by email?
Email ravi@citare.ai with “changelog subscribe” in the subject. One short email per week, no marketing spam.