AirQuery's embedded SDK lets your customers ask data questions in plain English — without ever leaving your app. It's not embedded charts. It's an embedded AI agent. Each tenant gets their own Thinkbox, their own data, their own brand.
For fifteen years "embedded analytics" meant putting a pretty chart of your customer's data inside your product. Useful but limited — they could see what you decided to show them, and nothing more. AirQuery rewrites the contract. Your customers ask, and a real agent answers — using their data, their metrics, their language. Inside your product. Under your brand.
The AirQuery agent slips into your product as a side panel, a modal, a button, or a full-page route. End-users ask in natural language; answers stream back, scoped to that customer's data, themed to your brand.
The host product (left) is your customer's view of your app. The chat panel (right) is AirQuery embedded — your colors, your name, your customer's data.
Every capability of the standalone Wise App, exposed as embed surfaces you can drop into any web app, mobile WebView, or extension.
Drop the <AirQueryAgent /> component into your app. Works with React, Vue, Svelte, plain JS, or via iframe for no-framework setups.
Your colors, your fonts, your logo. The agent looks like a native part of your product, not a third-party embed.
One AirQuery account, unlimited customer tenants. Each tenant sees only their own data, scoped automatically through their Thinkbox.
Sign a short-lived JWT in your backend with the tenant ID and permissions. The agent honours it. Works with Auth0, Clerk, Cognito, custom.
Per-user policies defined in the Thinkmap. Sales reps see their accounts, managers see their teams, admins see everything — enforced at the agent layer.
Each customer can have a Thinkbox tuned to their domain — their metrics, their ontology, their language — without one tenant ever seeing another's.
Token-by-token streaming, so answers feel instant. Reasoning, SQL, and chart all stream in progressively.
Prefer to render your own UI? The same agent is available as a REST/streaming API. Bring your own chat interface; we'll do the reasoning.
Pay per question, not per seat. Pass usage through to your customers or absorb it — your choice.
Install the SDK, mint a JWT for the current user on your server, drop the component into any page. That's it. You're shipping an AI analyst inside your product today.
Mobile? React Native SDK ships next quarter. iOS/Android WebView already works today.
You sell software. Your customers wish they could "see their data" inside it. Embed AirQuery and ship that capability without growing your data team. Charge for it as a premium tier.
You serve a specific industry (legal, healthcare, real estate). Your customers have rich domain data inside your product. Give them a domain-tuned AI analyst that speaks their language — without hiring ML engineers.
Sellers want analytics about their performance. Buyers want analytics about their procurement. Embed AirQuery on both sides with tenant-scoped Thinkboxes. One platform serves both views.
You build the internal data platform at a big company. Instead of every business unit asking for their own dashboard, embed AirQuery into the internal portal. Every team gets a self-service agent on day one.
The hardest part of embedded analytics isn't the embed — it's making sure tenant A can never see tenant B's data. AirQuery solves this at the agent layer, not just the SQL layer.
Each tenant's Thinkmap is a separate namespace. The agent cannot traverse outside it — not even if a prompt tries to trick it.
Every embed session is bound to a signed JWT with the tenant ID, user ID, and role. No JWT, no session. No exceptions.
RLS isn't applied to the SQL afterthought-style — it's a property of the Thinkmap that the agent enforces during reasoning, not just at execution.
Independent audit on policies, monitoring, encryption, access. Reports available under NDA.
Data residency in EU, US, or Singapore. PII handling controls. BAA available for healthcare customers.
Every embed question, with the tenant ID, user, timestamp, and DAF trace, streamed to your logging system (Splunk, Datadog, S3).
The embed SDK is included with every AirQuery Cloud plan. Free for < 1000 questions/month; usage-based after that. No per-seat fees, ever.
Request Embed Access Read the SDK Docs