Glossary

The Quri glossary

Plain-language definitions of the concepts and features behind Quri — what each one means and how to use it.

Core concepts

Proactive alerting

Proactive alerting means your tools tell you when a metric moves, instead of waiting for you to open a dashboard and notice. Quri watches the numbers you care about and pings you the moment signups, revenue, or a funnel step shifts — so you hear about problems before the standup does.

Anomaly detection

Anomaly detection is how Quri decides a metric has moved enough to be worth your attention. It learns each metric’s recent baseline, then flags the points that spike or drop well outside that normal range. That is the signal behind every proactive alert — a flagged change, not just a raw number.

AI analytics layer

An AI analytics layer sits on top of the tools you already run — PostHog, GA4, Razorpay, Stripe and more — and answers questions in plain language. It is not another dashboard to maintain. Quri reads across your connected sources, explains what changed, and alerts you, so the data comes to you instead of you digging for it.

ROAS (return on ad spend)

ROAS, or return on ad spend, is the revenue you earn for every unit of money you put into ads. You calculate it by dividing campaign revenue by ad spend, so a ₹40,000 return on ₹10,000 of spend is a 4× ROAS. Quri reads spend and revenue across your ad and payment sources and tells you when ROAS slips.

CPA (cost per acquisition)

CPA, or cost per acquisition, is what you pay in ad spend to win one customer or conversion. Divide total spend by the number of conversions and you have it: ₹20,000 spent for 50 signups is a ₹400 CPA. Quri tracks CPA per campaign across your connected sources and alerts you when it creeps up.

CTR (click-through rate)

CTR, or click-through rate, is the share of people who clicked your ad after seeing it. Divide clicks by impressions and multiply by 100: 200 clicks on 10,000 impressions is a 2% CTR. It tells you whether your creative earns attention. Quri reads CTR per campaign and flags a sudden drop before spend piles up.

CPM (cost per thousand impressions)

CPM, or cost per thousand impressions, is what you pay each time your ad is shown a thousand times. Divide spend by impressions, then multiply by 1,000. It measures the price of reach, not clicks or sales. Quri reads CPM per campaign across your ad sources so you can see when buying attention gets more expensive.

Attribution

Attribution is how you assign credit for a conversion across the touchpoints a customer met on the way — an ad, an email, a search, a return visit. Different models split that credit differently, which changes which channel looks worth the spend. Quri correlates events across your connected sources so credit follows the real path, not one platform’s claim.

First-party data

First-party data is the information you collect directly from your own users — page views, signups, orders, and product events on your site or app. You own it, so it stays accurate as third-party cookies fade. Quri Tag captures these events first-party and feeds them into your journey and metrics, with consent respected by default.

GEO (generative engine optimization)

GEO, or generative engine optimization, is the practice of shaping your content so generative AI engines surface and cite it in their answers. Instead of chasing a blue-link ranking, you write clear, self-contained, factual passages a model can lift and credit. Quri’s glossary is built this way — definitions stand alone as clean, citation-ready snippets.

AEO (answer engine optimization)

AEO, or answer engine optimization, is optimizing your content to be the direct answer an engine returns — the snippet a user reads without clicking through. You structure pages around real questions and give crisp, complete replies. Quri’s glossary pairs each term with a FAQ block so engines can pull a ready-made answer to follow-on questions.

AIO (AI optimization / AI overviews)

AIO covers optimizing for AI-overview surfaces — the AI-written summaries that now sit above search results and inside assistants. You earn a place in them by being clear, accurate, and easy for a model to summarise and trust. Quri publishes structured definitions and an llms.txt feed so these surfaces can read it cleanly.

MCP (Model Context Protocol)

MCP, the Model Context Protocol, is an open standard for connecting tools and data sources to AI clients like Claude or ChatGPT. A tool exposes typed actions over MCP, and the AI calls them safely. Quri ships a hosted MCP server, so you can ask your AI assistant about your metrics and Quri answers from your connected sources.

Cohort analysis

Cohort analysis groups users by something they share — usually the week or month they signed up — then tracks how each group behaves over time. It separates whether newer users stick around better than older ones, instead of blending everyone into one average. Quri reads your first-party events so you can compare cohorts without exporting raw tables.

Lookalike audience

A lookalike audience is a new group of people an ad platform builds to resemble a seed audience you supply — usually your best customers. The platform finds users with similar traits and behaviour, so you reach fresh prospects who look like proven buyers. Quri helps you define that seed from your own first-party customer data.

Time-to-contact

Time-to-contact is how long passes between a lead arriving and someone first reaching out to them. It is measured from the moment the lead lands to the first reply, and shorter is almost always better — interest fades fast. Quri can watch this gap and alert you when a lead has waited too long without a response.

Customer journey / identity spine

A customer journey is the full path one person takes across your tools — an ad click, a site visit, a signup, a purchase. An identity spine is the thread that ties those scattered events to a single person. Quri Atlas builds this spine from Quri Tag, so one user’s story reads end to end instead of splintering per tool.

AI analytics and anomaly detection

Sigma-based anomaly detection

Sigma-based anomaly detection flags a metric reading as unusual by measuring how far it sits from its own recent baseline, counted in standard deviations. Quri learns what a normal range looks like for each metric, then alerts you when a value drifts past that band, so a real spike or dip stands out from ordinary noise.

Segmented anomaly detection

Segmented anomaly detection runs the same checks per slice of your data — by city, channel, or cohort — instead of only on the overall number. A problem in one segment often hides inside a healthy total. Watching each slice on its own means a drop in one channel surfaces even when the headline metric looks fine.

Multi-source lagged-causal correlation

Lagged-causal correlation lines up a move in one source with a delayed move in another to suggest a likely cause. Ad spend today rarely shows up as revenue today; it lands a few days later. Quri checks for moves that track each other across that delay, so you get a plausible reason rather than a coincidence.

Free Tools hub

The Free Tools hub is a set of no-login calculators that let you try Quri-style checks on your own numbers in seconds. Spot an anomaly, find where a funnel leaks, work out a growth rate, or run a quick health check — all in the browser, with nothing to install, before you connect a single source.

Ads autopilot

Unified ads overview

A unified ads overview pulls your paid-ads numbers — spend, ROAS, CPA, CTR — into one screen across every connected ad account. Instead of logging into Meta and Google separately and doing the math in your head, you open Quri once and see how the money is actually working, in plain currency.

Multi-platform unified ads view

A multi-platform ads view puts Meta and Google side by side in one place, using the same columns for both. You compare ROAS and CPA across platforms without exporting two spreadsheets and lining them up by hand. When budget is tight, you can see at a glance which platform is carrying its weight.

Ad creative scoring

Ad creative scoring grades each live ad so you can tell, at a glance, which creatives are pulling their weight. Quri shows your running ads in a gallery and attaches an AI quality and performance score to each one, so you spend on the creatives that work and retire the ones quietly draining budget.

Ad spend anomaly alerts

Ad spend anomaly alerts tell you the moment your paid-ads spend breaks from its normal pattern — a campaign suddenly burning twice its usual budget, or returns falling off a cliff. Quri learns each account’s baseline, then pings you when spend strays past it, so a runaway ad does not quietly eat your week.

Guarded budget writes

Guarded budget writes let Quri change an ad budget — shift money toward a winner, or scale a strong campaign — but never on its own. Every change is prepared first, shown to you, and only runs after you confirm it. Quri then records who changed what and when, so nothing moves your money silently.

Guarded ad pause and resume

Guarded ad pause and resume lets you stop a runaway or underperforming ad — and switch it back on later — through Quri, with a confirmation step every time. Quri prepares the pause or resume, shows you exactly which ad it affects, and only acts once you approve, then logs the action for the record.

AI ads recommendations

AI ads recommendations are Quri’s plain-language “do this next” suggestions for your paid ads, each backed by the numbers it read. Rather than another chart to interpret, you get a short list of moves — scale this winner, trim that loser — with the evidence attached, so you act on what the data says instead of guessing.

Ads automation tactics

Ads automation tactics are opt-in rules that turn a signal into a recommended action — for example, when a campaign clears a return target, propose scaling it. You choose which tactics to switch on. They surface suggestions, not silent changes, so the writes they propose still pass through Quri’s prepare-then-confirm guard.

AI ad creative studio

The AI ad creative studio is where you generate and iterate ad creative inside Quri, without bouncing between separate design tools. Start from a prompt, get creative options, and refine them in place. Paired with creative scoring, you can shape new variants and judge how they stack up against the ads you already run.

Ads audience builder

The ads audience builder is where you define who sees your ads and expand reach with lookalikes. Build an audience from a seed, then ask for a lookalike to find more people who resemble it. It keeps audience work next to the creative and budget decisions, so you shape who, what, and how much in one place.

Competitor ad library

The competitor ad library surfaces the ads your rivals are running and helps you find who those rivals even are. Quri can auto-discover competitors and pull their live ads into one view, so you see the angles and offers they are testing — useful context before you brief a new creative or shift your own spend.

Customer journey (Atlas)

Customer-360 timeline

A customer-360 timeline is one chronological view of a single person across every tool you connect. Quri Atlas lines up their ad click, site visits, recorded session, signup, and purchase in order, so you read one story instead of cross-checking five dashboards to guess what the same user did.

Cross-source identity spine

The identity spine is how Quri decides two events belong to the same person. The Quri Tag drops a stable ID, then matches anonymous browsing to a known email once someone identifies. That link is what lets Atlas join a Meta click, a GA4 session, and a Shopify order into one real customer.

Landing attribution capture

Landing attribution capture records where a visitor came from the moment they arrive — the source, medium, and campaign on their first touch. Quri stamps that onto their profile so weeks later, when they buy, you still know which ad or post started it. First-touch credit survives even after the UTMs disappear.

User dropoff detection

User dropoff detection finds the stage where people quit your journey. Quri Atlas counts how many users reach each step — landed, viewed, signed up, purchased — and flags the gaps where the most fall away. Instead of suspecting checkout is leaky, you see the exact stage and how many you lost there.

Objective-aware funnel

An objective-aware funnel reads your journey through the lens of what your ads were actually trying to do. A campaign chasing purchases and one chasing signups should not be judged the same way. Quri Atlas weights the funnel by the dominant campaign objective driving traffic, so the steps you measure match the goal you set.

Session-recording playback

Session-recording playback lets you jump straight from a journey step into the video of what that person did. Saw a user stall at checkout on their timeline? Open the matching recording and watch the exact session — the clicks, the hesitation, the dead end. You move from a number to the real behaviour behind it.

GA4 BigQuery export

GA4 BigQuery export pulls your raw, per-event GA4 data into Atlas through BigQuery. The standard GA4 API hands back sampled, aggregated reports. The BigQuery export gives Quri one row per event, so the journey graph is built on what actually happened rather than GA4’s rounded-off summary of it.

PostHog identify email capture

PostHog identify email capture grabs the email a user passes when PostHog identifies them, then hands it to Quri’s identity spine. That email is the key that binds their earlier anonymous PostHog events to a known person, so a stream of anonymous activity finally becomes one named customer in Atlas.

Per-user events cron persist

Per-user events cron persist is the scheduled job that saves each person’s events into the Atlas journey store on a regular cadence. Rather than rebuilding everything live each time you open a profile, Quri keeps per-user history written down and current, so timelines and funnels load from stored data instead of a slow fetch.

Cold-tenant on-demand build

Cold-tenant on-demand build creates a journey graph the first time you open Atlas, before the scheduled job has ever run for you. New accounts have no stored history yet, so rather than show an empty screen, Quri builds the graph on the spot from your connected sources and renders it right away.

Competitive audit

Instant URL audit

An instant URL audit reads a website the moment you paste its address. Quri scans the page, pulls out what the brand sells and how it positions itself, and shows you a summary right away. The slower, deeper competitive crawl runs in the background and is ready by your next visit.

Own-site synthesis

Own-site synthesis is Quri reading your own website and writing a clear profile of your brand from it. It pulls what you sell, who you sell to, and how you describe yourself, then turns that into the baseline every competitor comparison is measured against — no questionnaire, no manual setup.

Auto-discover competitors

Auto-discover competitors finds your rivals for you. From your own site profile, Quri surfaces five category leaders and five direct peers — the big names you measure against and the close ones you fight for the same customers. You do not have to name a single competitor first.

Deep competitor crawl and domain caching

A deep competitor crawl reads a rival’s pages in depth — not just the homepage — to learn how they sell. Quri caches each domain it crawls, so the next audit that touches the same competitor reuses the saved read instead of fetching it again. That keeps audits fast and cheap.

Page-tactic detection

Page-tactic detection spots the moves a web page makes to convert and rank. Quri reads a competitor’s page and flags the conversion plays, SEO signals, and messaging angles it uses — social proof, pricing framing, calls to action, and more — so you can see exactly how rivals are trying to win.

Evidence-backed recommendations

Evidence-backed recommendations are audit suggestions tied to what Quri actually crawled. Every recommendation points back to the competitor page or your own profile that prompted it, so you can see why Quri is suggesting it. No vague best-practice filler and no invented numbers — just findings grounded in real pages.

Per-plan crawl caps

Per-plan crawl caps set how deep and how wide an audit can crawl based on your plan. Higher plans crawl more competitors and more pages each; smaller plans get a tighter, focused read. The caps keep crawls predictable and costs in line with what each plan includes — no surprise overruns.

Cross-tenant domain reuse

Cross-tenant domain reuse means when one customer’s audit crawls a public competitor, Quri can serve that saved read to another customer’s audit of the same public domain. It only ever reuses public-page reads — never your private metrics or account data — which cuts crawl cost and makes repeat audits faster.

Instagram

Instagram Marketing Analytics

Instagram marketing analytics turns your account activity into clear next moves. You see how reach, views, and engagement shift over time, which posts pull their weight, and where you are losing ground. Quri reads your business account and tells you what changed and what to do about it.

Instagram Business Account

An Instagram business account (also called a professional account) unlocks the insights and discovery data that personal accounts hide. Personal profiles return no analytics at all. You convert in the Instagram app under account settings, then connect to Quri to read reach, views, and per-post metrics.

Instagram Audit

An Instagram audit reviews your profile against the habits that grow accounts: posting cadence, hook quality, engagement, and bio clarity. Quri grounds every finding in your real data, then ranks fixes by impact so you start with the change that earns the most, not a generic checklist.

Reach vs Views on Instagram

Reach counts the unique accounts that saw your content. Views count total plays, so one person can add several views. Instagram now reports views as the unified play metric and no longer exposes impressions. Reading both tells you how wide you spread and how often people watch.

Instagram Engagement Rate

Engagement rate measures how much your audience reacts, dividing interactions like likes, comments, saves, and shares by reach or followers. It tells you whether content lands, not just how far it travels. Quri tracks the rate over time so you catch a slide before it drags your growth down.

Instagram Competitor Analysis

Instagram competitor analysis studies the public accounts you compete with: their posting cadence, captions, formats, and which posts draw reactions. Be clear on the limit: Instagram only exposes public likes and comments for accounts you do not own. Reach, saves, shares, and demographics stay hidden for competitors.

Business Discovery API

Business Discovery is the Instagram API that returns public data about other professional accounts by username. It gives you a competitor’s profile basics and recent posts with like and comment counts, captions, media type, and permalinks. It cannot return reach, saves, shares, or any audience demographics.

Meta Ad Library Competitor Tracking

The Meta Ad Library is a public archive of active ads across Facebook and Instagram. You can see a competitor’s live ad creatives, copy, start dates, and platforms. Exact spend stays hidden for ordinary commercial ads, so treat ad longevity, not budget, as the signal that something is working.

Reels Hook Analysis

A hook is the first few seconds of a reel that decides whether someone keeps watching. Hook analysis studies those openers across your reels and your competitors to find the patterns that hold attention. Quri reviews captions and thumbnails by default and can render public reels for deeper review when you opt in.

Instagram Anomaly Alerts

Anomaly alerts watch your Instagram metrics and tell you when something breaks from its normal pattern, like a reach collapse or an engagement spike. You do not have to check a dashboard every day. Quri runs the detection on a schedule and messages you only when a change deserves attention.

AI chat

Conversational analytics

Conversational analytics lets you ask about your stack in plain language and get a grounded answer back. Instead of building a query or hunting through dashboards, you type a question into Quri chat and it reads the right metrics across your connected tools, then replies with the number and where it came from.

Live connector reads

Live connector reads mean Quri chat answers from your current data, not a cached snapshot from last week. When you ask a question, it queries the connected source — PostHog, GA4, Razorpay and the rest — at that moment and cites which one it read, so you can trust the number is fresh.

Real-time streaming

Real-time streaming means Quri chat writes its answer out word by word as it thinks, instead of making you wait for the whole reply at once. You see the response take shape immediately, so a longer answer never feels like a frozen screen, and you can start reading before it finishes.

Conversation history

Conversation history keeps your past chats with Quri so you can come back to them. A question you asked last week, and the answer Quri gave, stay saved and revisitable. You pick up where you left off instead of re-explaining context, and you have a record of what the numbers were when you asked.

Cost per conversation

Cost per conversation shows what each Quri chat actually costs to run. Every answer uses a language model, and that has a real price. Quri tracks the model cost of a conversation and makes it visible, so you can see which questions are cheap and which heavy ones add up, with no hidden meter.

Controllable time window

A controllable time window lets you scope a chat answer to the date range you care about. Ask about the last 7 days, last month, or a specific stretch, and Quri reads only that period. You control the window, so an answer about “this week” means your week, not a default the tool chose for you.

Smart context trimming

Smart context trimming keeps a long conversation inside the model’s budget without dropping what matters. As a thread grows, Quri decides which earlier turns and data to carry forward and which to set aside. You keep asking follow-ups in one thread, and the answers stay relevant instead of breaking on length.

Error and tool-failure surfacing

Error and tool-failure surfacing means that when a read or a tool fails, Quri chat tells you plainly instead of inventing a number. If a connector times out or a source is unreachable, you see what went wrong rather than a confident-looking but wrong answer. Honesty about gaps is the whole point.

Smarter-cheaper model selection

Smarter-cheaper model selection routes each chat turn to a model that fits the job. A simple lookup does not need the heaviest, priciest model, while a hard reasoning question does. Quri picks per turn, so you get a capable answer without paying top rates on every question, and the cost stays proportional to the work.

Knowledge-graph grounded answers

Knowledge-graph grounded answers let Quri chat reply from a connected map of your metrics, users, alerts and the links between them. It reads that graph first for a fast, cheap answer, falls back to a live source only when a slice is missing or stale, cites the exact records it used, and never invents a number.

Semantic graph search

Semantic graph search lets Quri chat find the right emails, chat summaries and recommendations by meaning, not exact words. It turns your question into a vector, compares it against everything Quri has written down, and returns the closest records — so "what did we tell users about the checkout bug?" surfaces the message even if it never said "checkout".

Private embeddings

Private embeddings are how Quri turns your text into the vectors behind semantic search. By default it uses a local model, so your emails, chat summaries and recommendations never leave your environment and there is no per-word cost. You can switch to a managed embedding service later, but the private default needs no key.

AWS-hosted models

AWS-hosted models let Quri run the same Claude models through your AWS account instead of a direct API key. You get IAM-based access, unified AWS billing and data residency in your region, with no change to the answers chat gives. It is an opt-in host swap — the direct path stays the default.

Segmentation

Geographic segmentation

Geographic segmentation splits any metric by where it happens — country, region, or city — so a single national number becomes a map of where things actually move. Instead of reading one blended figure, you see which places drive revenue, signups, or ad spend, and which are quietly dragging the average down.

Revenue by city cohorts

Revenue by city groups your earnings into per-city cohorts so you see where the money truly comes from, not just a national total. Quri ranks cities by the revenue they contribute, so you can tell whether one metro is carrying the business and where a small push might turn a promising city into a real one.

Segment opportunity detection

Opportunity detection reads every segment and tells you what to do with it — scale the ones outperforming, fix the ones lagging, or hold the balanced ones steady. Rather than scanning a long table and guessing, you get each segment tagged with a recommended move, so your attention lands on the slices that are worth acting on.

Dimension switcher

The dimension switcher lets you change how a metric is broken down on the fly — country one moment, device the next, then campaign — without leaving the view or rebuilding a report. You follow a number wherever your question leads, asking "where", then "on what", then "from which campaign" against the same metric and window.

Custom segment labels

Custom segment labels let you name and save the groupings that matter to your business — "tier-1 metros", "trial cohort", "high-intent regions" — instead of re-deriving them every time. Once saved per workspace, a label is reusable across views, so the team reads the same named segments and stops arguing over what each slice actually contains.

Per-workspace breakdown config

Per-workspace breakdown config sets which dimensions a workspace tracks by default — geography for one team, device and campaign for another. Each workspace tunes its own cuts, so the segment views and the dimension switcher offer the breakdowns that team actually uses, instead of burying everyone under every possible slice.

Segment proactive alerts

Segment proactive alerts watch each slice on its own, not just the headline number. When one city, device, or cohort moves abnormally — even while the global metric looks calm — Quri flags that segment specifically. A problem hiding inside an otherwise flat total surfaces early, so you catch the city that cratered before it drags the rest down.

Held-segment revenue

Held-segment revenue is the revenue Quri can attribute to a specific audience or cohort it has identified — a region, a custom label, a high-intent group — and surface as its own line. Rather than a blended total, you see what a particular segment is worth, which makes the case for doubling down on it concrete.

Connectors

Connector

A connector is a swappable adapter that reads one of your tools — Stripe, Meta, GA4, PostHog, Shopify — and maps its data into the shape Quri understands. Add a source by connecting it; you never touch the rest. Each connector handles its own API quirks so every metric lands in one consistent format.

OAuth federation

OAuth federation lets you connect a source with one click instead of pasting API keys. You sign in to the tool, approve what Quri may read, and the source returns a token. No secret is ever copied by hand, and you can revoke Quri from inside that tool whenever you choose.

Metric discovery

Metric discovery is how Quri finds the metrics and dimensions a source offers, so you do not have to name them yourself. Once a tool is connected, Quri asks it what data exists — revenue, sessions, spend, breakdowns — and lists what you can read. You pick from a real catalog instead of guessing field names.

Currency normalization

Currency normalization makes money from different sources comparable. Quri reads each value as an ISO-4217 currency in its smallest unit — cents, paise, yen — and stays aware that some currencies have zero or three decimals. So a Stripe charge, a Shopify order, and Meta ad spend add up correctly instead of mixing units silently.

Token-refresh automation

Token-refresh automation keeps your connections alive without you logging back in. OAuth access tokens expire fast — often within an hour. Quri stores the refresh token and quietly trades it for a fresh access token before the old one dies, so a source you connected weeks ago keeps syncing instead of silently going dark.

Disconnect and data purge

When you disconnect a source, Quri stops reading it and deletes the data it pulled from that source. Disconnecting is the off switch and the cleanup in one step: the token is dropped so no further reads happen, and the metrics Quri stored from that connector are purged, supporting your GDPR obligations.

Per-workspace encrypted credentials

Every credential you connect is encrypted at rest and scoped to one workspace. A token belongs to the workspace that added it and is never readable by another tenant. So when two companies both connect Stripe, their keys stay sealed and separate — Quri scopes access by construction, not by a filter someone could forget.

Rate-limit handling

Rate-limit handling keeps a busy source from breaking your whole sync. Every external API caps how often you may call it. When a connector nears that cap, Quri backs off and retries rather than hammering the source, and skips what it cannot read this round instead of failing everything — degrading gracefully so the rest still updates.

Alerts and notifications

Alert channels

Alert channels are the places Quri can send an alert when a metric moves — email, Slack, a webhook to your own system, or a Linear ticket. You pick where each alert lands, so a revenue drop can hit Slack while a routine flag goes to email. One detected change, delivered wherever your team already watches.

Severity-based routing

Severity-based routing sends each alert to a different place depending on how serious it is. A critical drop can fire to Slack and email at once, while a low-priority flag quietly joins a digest. You set the rules once, so urgent problems interrupt you and minor ones wait, instead of every alert shouting equally.

Digest batching

Digest batching collects your lower-priority alerts and sends them together on a schedule, instead of pinging you for each one. A handful of small flags arrive as a single summary you can scan in a minute. Critical alerts still fire instantly — the digest is for the noise that does not need an interruption.

Per-workspace channels

Per-workspace channels let each workspace set its own alert destinations, kept separate from every other. One workspace can route alerts to its own Slack while another uses email and a webhook, with no crossover. If you run several brands or clients, each one gets its alerts where its team watches, never mixed together.

User preferences and opt-out

Notification preferences let each person choose what they hear about and how. You can mute channels you do not need, pick which alert types reach you, and opt out of non-essential messages entirely. Quri honours that choice on every send, so people stay informed about what matters to them without being buried in the rest.

Lifecycle emails

Lifecycle emails are the product messages Quri sends as you use it — a welcome when you join, a nudge when a useful step is still undone, a check-in if you go quiet. They are timed to what you are doing, not blasted on a schedule, so each one arrives when it is genuinely worth reading.

Auth emails

Auth emails are the security-critical messages tied to your account — verifying your address, resetting a password, or confirming a multi-factor step. Quri sends these through Relay so they go out the same trusted way as your other email. Because they protect access, they always send regardless of your other notification preferences.

Comms audit log

The comms audit log is the record of what Quri sent, to whom, and when. Every alert, lifecycle message, and auth email leaves an entry, so you can answer “did this person get notified?” with evidence rather than a guess. It turns delivery from a black box into something you can check after the fact.

SES feedback suppression

SES feedback suppression keeps your email reputation healthy by listening to what the mail provider reports back. When an address hard-bounces or marks a message as spam, Quri stops sending to it. Continuing to email dead or hostile addresses drags down deliverability for everyone, so suppression protects the inbox placement of the people who do want your alerts.

Per-tenant consent-log gate

The consent-log gate checks a per-tenant consent record before any message leaves. If a tenant has not granted consent to be contacted on a channel, Quri holds the send rather than risking an unwanted message. It is a delivery guard built into Relay 0.3.0, so compliance is enforced at the point of sending, not bolted on later.

Social watch alerts

Social watch alerts let Quri keep an eye on what people say about your brand and rivals on the open web without you logging in. In the background it tallies daily mentions and reads their mood, then pings you when chatter suddenly spikes or collapses, or when sentiment takes a sharp turn negative across enough posts to matter.

Watched terms

Watched terms are the names you tell Quri to keep an eye on across the social web — your brand, your competitors, and the topics in your category. Quri tracks mentions of each one and reads the mood behind them, so the background social watch knows what to listen for and which buzz is worth pinging you about.

Anomaly social enrichment

Anomaly social enrichment answers the question every metric alert leaves hanging: why. When a number moves — revenue drops, signups spike — Quri pulls the matching social chatter from the last month and attaches it to the alert. So instead of just “revenue down 30%”, you get the threads where people are saying why.

Onboarding

Guided setup wizard

The guided setup wizard is the first-run flow that takes a brand-new Quri workspace from empty to useful. It walks you through the basics, helps you connect your first data source, and gets you to a real insight in a few steps, so you reach value without guessing what to set up or in what order.

Pulse checklist

The Pulse checklist is the progress tracker on your Quri home screen. It shows which setup steps you have done and which remain, and it nudges the next step worth the most to you right now. Instead of a flat to-do list, it points you at the highest-value action so early effort pays off fast.

Website scan during onboarding

A website scan during onboarding lets you paste your URL while setting up Quri. Before you connect any data source, Quri reads your site to learn what your product is, who it serves, and how it is positioned. That context seeds your workspace, so early answers and audits land closer to your business from the start.

Demo taste step

The demo taste step lets you try Quri on sample data with zero connections. Before you hand over access to a real source, you can ask a question and see a grounded answer come back, so you feel what Quri does first. It turns the connect ask from a leap of faith into a confirmed yes.

First-visit intro cards

First-visit intro cards are the short, in-context cards Quri shows the first time you land on a surface. Rather than a wall of docs, each card explains what this screen does and what to try next, right where you are. They orient a brand-new user fast, then step aside once you know your way around.

Behavioral checklist completion

Behavioral checklist completion means Quri marks a setup step done when you actually do it, not when you tick a box. Ask Chat a real question and the "try Chat" step completes on its own. The checklist tracks real usage, so your progress reflects what you have genuinely done rather than a list you clicked through.

Workspaces and teams

Workspace switcher

The workspace switcher lets you move between the accounts you have access to without logging out and back in. Agencies and teams jump from one client to the next in a click. Everything you see — metrics, alerts, connectors — reloads scoped to the workspace you picked, so you never mix two clients up.

Workspace hard-delete

A workspace hard-delete permanently removes a workspace and the data scoped to it. When you stop working with a client, you delete their workspace instead of leaving it idle. The action is deliberate and asks you to confirm, because once it runs the workspace and its stored data are gone, not merely hidden.

Cross-tab sync

Cross-tab sync keeps the workspace you picked and your signed-in state consistent across every browser tab. Switch a client in one tab and your other open tabs follow, so you never act on stale data. Sign out anywhere and every tab signs out too, which keeps a shared computer from leaking the last account.

Auto-scoped data

Auto-scoped data means every query and every stored credential is tied to the active workspace by construction, not by a filter someone has to remember to add. There is no path that reads across tenants, because the scope is built into how Quri stores and fetches. One client can never see another client’s numbers.

Workspace settings

Workspace settings is where you tune one workspace to match a client — their home country, reporting currency, and the breakdown dimensions you slice metrics by. Each setting applies to that workspace alone, so a client in India and one in the US each read their own currency and region without you reconfiguring between switches.

Agency multi-subaccount

Agency multi-subaccount lets one agency account hold many client subaccounts under it. You manage every client from a single login, switching between their workspaces as you go. Each subaccount keeps its own connectors, settings, and data, so the agency gets one front door while each client stays cleanly separated from the rest.

Per-subaccount entitlements

Per-subaccount entitlements scope a plan and its capabilities to each subaccount on its own. One client can be on a higher tier with more connectors while another stays lean. The agency above them does not force a single plan down to every client, so each subaccount carries exactly the features its work needs.

Orgs and roles

Orgs and roles map the people on your team to what they may do. An organization groups your users, and a role on each user decides their permissions — an admin manages connectors and billing while an analyst reads metrics. Roles drive the capabilities Quri grants, so access matches the job rather than handing everyone the keys.

Plan-based feature gating

Plan-based feature gating decides which features a workspace can reach based on its plan. Quri has three tiers — free, pro, and scale — and each maps to a set of capabilities. Move up a tier and more unlocks; features above your plan stay locked until you upgrade. Gating is enforced on the server, not hidden in the UI.

Fine-grained capabilities

Fine-grained capabilities are the named permissions that sit under roles and plans. Each action Quri can take — manage connectors, view a waitlist, run a write — checks for a specific capability. The model is default-deny: anything not explicitly granted is refused. So an undeclared surface is locked by default, never quietly left open.

Quota enforcement

Quota enforcement caps how much a workspace can do over a window — for example, the write actions allowed per day. Each plan tier sets the ceiling, and Quri counts usage against it per workspace. Once a workspace hits its limit, further actions wait or are refused until the window resets, so no tenant runs away with shared capacity.

LLM cost quotas

LLM cost quotas cap how much AI spend a workspace can run up. Chat answers and AI features call language models, and that costs money per request. A cost quota puts a ceiling on it per tenant, so a heavy month in one workspace stays within bounds instead of producing a surprise bill that lands on the whole account.

Per-tenant consent export

Per-tenant consent export hands you a workspace’s recorded consent decisions — who allowed tracking and who opted out. When a client asks for proof or a regulator wants records, you export that workspace’s consent data on its own. The export is scoped to one tenant, so a client only ever receives their own records, never another client’s.

Auth and security

Email and password signup

Email and password signup is how you create a Quri account without a third-party login. You enter your work email and a password, Quri sends a confirmation code to that inbox, and you type it back to prove the address is yours. Only then does the account go live and ready to use.

OAuth signup with federation

OAuth signup lets you join Quri with an identity provider you already trust instead of inventing another password. You authorize Quri through that provider, it confirms who you are, and Quri creates your account from the verified identity it hands back. There is no separate password for Quri to store or for you to forget.

Invite-only waitlist gating

Waitlist gating keeps signup invite-only. When it is on, Quri checks your email against the approved waitlist before any account is created. Approved addresses sail through; everyone else is asked to join the list and wait for an invite. It lets Quri onboard accounts in a controlled order rather than opening the doors to all.

Resend confirmation code

Resend confirmation gives you a fresh code when your first one expired or never landed. Confirmation codes are short-lived on purpose, so a slow email or a closed tab can leave you stuck. Rather than restart signup, you ask Quri to send a new code to the same address and finish verifying from where you left off.

Custom login page

The custom login page is Quri’s own branded sign-in screen rather than a generic hosted one. It signs you in and, when your account needs an extra step, walks you through it in place — setting a new password, or entering a one-time MFA code. You stay on Quri instead of bouncing to a third-party UI.

Password reset

Password reset gets you back in when you forget your password. You ask Quri to start a reset, it emails a one-time code to your address, and you use that code to set a new password. Quri returns the same neutral response whether or not the email has an account, so a reset request never leaks who is registered.

Account settings

Account settings is where you manage your Quri identity in one place — your profile details, the workspace you belong to, and your security preferences. From here you update how you appear, confirm which workspace you are in, and reach the controls that govern sign-in, connected sources, and the tokens that authorize guarded actions.

GDPR data-deletion callback

The data-deletion callback is the compliance endpoint that honors a deletion request from a provider or a user. When someone exercises their right to be forgotten — through a connected platform’s data-deletion flow — that request reaches Quri at this endpoint, and Quri removes the matching data for the person it names. It is how deletion requests stay honored automatically.

Hard account delete

Hard account delete permanently removes your Quri account and the data tied to it. This is not a pause or a hidden archive — it is the irreversible option for when you are done with Quri for good. Because it cannot be undone, Quri makes you confirm before it runs, so the choice to wipe everything is always deliberate.

Write-token management

Write-token management is where you mint, list, and revoke the tokens that authorize guarded write actions, like issuing a refund. Each token is minimum-scope, shown once at creation, stored encrypted with KMS, and every use is audited. You hand out only the power an action needs, see what is live, and revoke any token the moment you stop trusting it.

OAuth auto-workspace binding

OAuth auto-workspace binding puts a federated user in the right workspace on their first sign-in, with no manual step. When you sign in through a provider, Quri reads your verified identity and attaches your account to the workspace it belongs to, so you land where your team already works instead of in an empty account to wire up yourself.

Source auto-connect

Source auto-connect lets you wire up a data source as part of, or right after, signing in. Instead of leaving a fresh account staring at an empty screen, Quri guides you to connect a source early so it has metrics to read from day one. You go from sign-in to a connected, working account in one continuous flow.

Performance and platform

Stale-while-revalidate

Stale-while-revalidate serves you the last cached answer instantly, then refreshes it in the background. Your screen fills right away with recent numbers instead of spinning on a cold fetch, and the fresh value swaps in moments later. You see a freshness banner so you always know how recent the figure on screen is.

Manual cache purge

Manual cache purge is the refresh button for your data. Caching keeps reads fast, but sometimes you change something at the source and want the newest figure now. Purging clears the cached value for that read, so the next load fetches straight from the source instead of serving what Quri remembered a moment ago.

Data freshness banner

The data freshness banner tells you how recent the number on screen is. Because Quri caches reads to stay fast, a figure might be a few minutes old — so a small banner shows its last fetch time and gives you a refresh control. You always know if the figure on screen is cached or the latest.

CDN edge caching

CDN edge caching keeps a copy of responses close to you on a content network, so requests do not always travel to the origin server. The result is faster page loads and less load on Quri itself. The cache is keyed by who you are, so one workspace never serves its cached response to another.

Cache prewarming

Cache prewarming fills Quri’s caches and metric catalogs ahead of time, so your first read of the day is not a cold one. Instead of waiting for you to ask before fetching, Quri populates the common queries in advance. When you open a view, the answer is already warm and loads fast rather than rebuilding from scratch.

Warm-keys read path

The warm-keys read path routes your common queries through a set of cache keys that Quri keeps populated. When you ask for a metric people read often, the answer comes from a warm key instead of being recomputed each time. You get a fast response, and Quri spends less work and cost rebuilding the same figure repeatedly.

DynamoDB reaping

DynamoDB reaping auto-expires old rows so Quri’s store stays lean. Records that no longer need to live — short-lived caches, aged chat transcripts — get a time-to-live stamp, and the database deletes them automatically once that time passes. You keep the data that matters and pay nothing to store rows that have outlived their use.

Async job queue

An async job queue moves heavy work off your request so the screen never hangs waiting. When Quri has a big sync or rebuild to do, it hands the job to a queue and answers you right away. A worker picks the job up, an idempotency key stops duplicates, and a second worker adds throughput and reliability.

Chat cost cancellation

Chat cost cancellation lets you stop an answer mid-stream. When you ask Quri a question and realize the reply is going the wrong way, you hit Stop. The in-flight model call is cancelled, the tokens it would have spent are not billed, and you are free to rephrase — so a misfired question does not run up cost.

Per-call cost tracking

Per-call cost tracking records the cost of every model call Quri makes on your behalf. Each chat answer carries a small token cost, and Quri logs it precisely rather than rounding it away. That gives you an honest, itemized picture of what your AI usage costs, so spend stays transparent instead of arriving as one opaque number.

Product context and enrichment

Full-page website scrape

A full-page website scrape reads more than your homepage. Quri crawls your public pages — features, pricing, about — to build a rounded picture of what you sell and how you say it. That fuller context is what lets Quri answer questions and audit your site without you typing your product details by hand.

JS rendering for SPA pages

Many product sites are single-page apps that ship an empty shell and draw their content with JavaScript. A plain fetch reads that shell as blank. JS rendering runs the page in a headless Chromium lambda so the real content loads first, then Quri reads it. If rendering fails, it falls back to a plain fetch.

Scrape progress UX

Scrape progress UX shows you the crawl happening rather than a frozen spinner. As Quri fetches your homepage and then works through your other public pages, the scan surface reports what it has read and what is still pending, so you know enrichment is moving and roughly when your full context will be ready.

Thin-page skip

Thin-page skip keeps near-empty pages out of your product context. When a crawled page carries almost no real content — a bare redirect, a loading shell, a cookie wall — Quri detects it and leaves it out instead of feeding noise into the knowledge graph. Your context stays built from pages that actually describe your product.

Resilient field extraction

Resilient field extraction pulls the fields that matter — your product name, what it does, features, and pricing signals — even out of messy or inconsistent markup. Real sites are rarely clean, so Quri reads the meaningful content rather than relying on tidy tags. When a field genuinely is not there, Quri marks it not detected instead of inventing one.

DNS-rebinding protection

DNS-rebinding protection stops a scrape from being tricked into reaching internal addresses. Before Quri fetches a URL, it checks where that host actually resolves and blocks anything pointing at private or internal ranges. This shuts down SSRF and DNS-rebinding tricks, so a crafted target can never make the scraper hit your network or ours.

Product knowledge graph

The product knowledge graph is the structured set of facts Quri holds about your product — what it is, who it serves, its features and pricing signals — connected so Quri can reason over them. It is built from your scraped pages and inferred by a language model, with low-confidence or hallucinated links dropped so the graph stays grounded.

Chat context tool

The chat context tool lets Quri chat read your captured product context while it answers. Instead of replying from generic knowledge, chat pulls the facts in your knowledge graph — what you sell, your features, your pricing — and grounds its answer in them. That is why Quri talks about your product specifically rather than products in general.

Context-aware budget floor

A context-aware budget floor sets a sensible minimum for an ad budget using what Quri knows about your product, instead of a blind one-size default. By reading your captured context — your category and pricing signals — Quri lands on a floor that fits your business, so a budget recommendation starts from a grounded number rather than an arbitrary one.

Login-walled context capture

Login-walled context capture reads pages that sit behind a login, but only when you authorize it. Some product detail lives inside an account or a gated docs area that a public crawl never sees. With your permission, Quri captures that context too, so its understanding of your product is not limited to what an anonymous visitor can reach.

marketing-advisory

CRO diagnosis

CRO diagnosis reads the conversion funnel Quri already holds and finds where it leaks. It checks each step against researched benchmarks — checkout abandonment, the MQL-to-SQL handoff, per-step floors — and flags the ones falling short. For every issue you get a plain finding and a short playbook of what to try first, not generic advice.

Churn diagnosis

Churn diagnosis reads your retention curve and judges it by shape, not by a single number. A curve that flattens and holds means product-market fit is present; one that keeps falling past month three points to a fit problem, not onboarding; a curve that dips then climbs is recovering. Each shape comes with its own retention playbook.

Ad fatigue diagnosis

Ad fatigue diagnosis scores each creative against its own baseline with one composite index — click-through decay, conversion drop, rising CPM, and climbing CPA. Falling clicks warn early, before ROAS slides. Quri flags fatiguing creatives, saturated audiences, and exhausted reach, then hands you a playbook of what to refresh first. It reads creative; it never writes it.

Pricing diagnosis

Pricing diagnosis compares your price and tiers against the competitor prices Quri already crawls. It flags when you sit far from where rivals cluster — priced high needs a visible reason, priced low leaves money behind — and when you offer fewer tiers than the norm. It is guidance to check, and never writes copy.

Proactive marketing advice

Proactive marketing advice is the recommended fix Quri attaches to an alert the moment a metric breaks. When a conversion, retention, ad, or pricing number moves, it runs the matching diagnostic on the spot, picks the most serious finding, and ships the playbook of what to try inside the alert itself.