01
Compose
The model chooses registered primitives and fills structured props. No markup, no CSS, no frontend code.
AI answers,rendered as product UI.
Leara gives models a safe interface contract, then renders the result with your app's own components, state, and theme.
Public access opens September 27, 2026 in 25d 9h 14m
Generated product UI
Revenue report rendered by your app
No generated code
Applied
Revenue
$284.5M+12.4%YoY
Pipeline
$118.6M+18.1%
Win rate
46.3%+0.4%pp
CAC
$188-7.2%
Revenue momentum
Monthly booked revenue and pipeline growth
- Revenue
- Pipeline
Sales by segment
Share of revenue by customer segment
- Enterprise38.0%
- Mid-market27.0%
- Small business18.0%
- Consumer12.0%
- Public sector5.0%
Channel performance
Revenue quality by acquisition channel
3 rows
| Channel | Revenue | Pipeline | Win rate | Status |
|---|---|---|---|---|
| Enterprise salesField + strategic | $82.4M | $48.0M | 52.1% | Scaling |
| PartnerCo-sell + referrals | $54.6M | $29.4M | 49.8% | Healthy |
| Paid acquisitionSearch + social | $68.1M | $31.2M | 38.7% | Watch |
Key insights
What the operating signals suggest
3 insights
Revenue quality is improving
- Enterprise and partner channels carry the strongest revenue quality.
- Pipeline is expanding faster than booked revenue.
- Paid acquisition should stay on watch while CAC normalizes.
?
show Q2 revenue risk
Runtime loop
The model composes.Leara validates.Your app renders.
One request becomes native UI your customers can use. The model names what to show, Leara enforces the contract, and your product owns the final surface.
Model/Runtime/Application
02
Validate
Leara checks every component name, prop, and tree shape before anything reaches your UI.
03
Render
Your app turns the tree into real product surfaces with your components, theme, state, and permissions.
Built for every use case
Leara helps AI products turn responses into interfaces users can explore, share, and act on.
AI Knowledge Search
Surface precise answers from docs, websites, and knowledge bases with cited summaries, cards, and suggested next steps, no scrolling through links.
Summarize the reimbursement policy for international travel and show approval limits
Reimbursement policy for international travel
Reviewed the latest HR policy documents, regional employee handbooks, and leave compliance guidelines across US, EU, and APAC regions.
RegionBudgetMax carryoverApproval required
United States$400/day40 hoursYes
European Union$350/dayup to 20 daysYes
United Kingdom$350/dayup to 10 daysNo
China$200/day1 weekYes
South-east Asia$125/dayup to 20 daysNo
India$125/day1 weekNo
LearaComposeProduct proof
One generated interface, rendered wherever the user is working
Start with a real user question. Leara returns a validated interface tree, then your product decides whether it replaces the current screen or appears as a drawer beside the native workflow.
Inline replacement
route.ts
1import { compose } from "liara";23export async function POST() {4 return await compose({5 prompt: "Show renewal risk for enterprise accounts this quarter",6 primitives: [Metrics, Chart, DataTable, Summary],7 surface: "native-view",8 theme: "system",9 });10}1112// validated in 41ms - 4 registered primitives
Generated where work happens
Replace the current view or open a drawer without sending users somewhere else.
Invalid output stops early
Unknown components, bad props, and malformed trees never mount.
Looks like your product
Generated surfaces inherit the primitives and styling users know.
Adapts to the surface
The same prompt can become an inline dashboard or a contextual side panel.
Provider agnostic
Use the model stack your product already trusts.
Leara gives each provider the same compact interface contract, then validates the output before your app renders it. The same boundary is designed for the model stack your team already trusts.
OpenAI
GPT models
Anthropic
Claude
Google Gemini
Gemini
Mistral
Mistral
xAI
Grok
Cohere
Command
Getting started
Start with Leara primitives. Customize the product surface.
Use Leara's built-in primitive kit as the interface vocabulary, customize the contract around your product, then let the runtime validate, render, and reuse the result.
A practical path from first primitive to production-ready generated interface.
01
Register
Start with Leara's own primitives for the product surfaces AI should be allowed to create.
MetricChartDataTableSummaryForm
02
Request
Pass the model a prompt, your customized primitive set, and the surface where the answer should appear.
PromptPrimitivesSurface
03
Validate
Leara checks component names, props, schema shape, and permissions before anything can render.
NamesPropsSchemaPermissions
04
Render
Leara validates the tree, then your product renders it as a dashboard, drawer, report, or workflow.
ComponentsStateThemeRouting
05
Reuse
The same contract can power dashboards, chat drawers, records, reports, and internal workflows.
DashboardsDrawersReportsWorkflows
Developer experience
Built to feel like infrastructure, not a prompt demo.
Register what your product can do, let the model compose within those boundaries, then render through your own design system.
liara.config.ts
1import { defineLiaraConfig } from "@liara/core";2 3export default defineLiaraConfig({4 endpoint: "/api/liara/compose",5 theme: {6 font: "Geist",7 radius: "12px",8 accent: "blue"9 },10 primitives: [11 Section,12 StatGrid,13 EntityTable,14 Trend,15 RelatedQueries16 ]17});Pricing
Start with the runtime. Pay when it becomes product traffic.
The SDK stays easy to try. Production usage is metered around the generated UI outputs your product serves.
Free
$0/ month
For local development, SDK trials, and first product tests.
Included100 UI outputs
Requests300 requests
most common
Developer
$29/ month
For solo builders validating AI-native application surfaces.
Included25k UI outputs
Requests100k requests
Team
$149/ month
For product teams embedding Leara inside production SaaS apps.
Included250k UI outputs
Requests1M requests
FAQ
Questions before you put generated UI in production.
The important boundary is simple: the model proposes structure, Leara validates it, and your product stays in control.
What is Leara?
Leara is a runtime for AI-native product surfaces. Models choose semantic interface primitives, Leara validates the result, and your application renders the final UI.
Is this another AI SDK?
No. AI SDKs help you call models. Leara is the layer after model output: composition, validation, rendering contracts, observability, and usage controls for generated product UI.
Does the model generate React or HTML?
No. Models never generate HTML, JSX, CSS, or executable UI code. They compose a constrained Interface that your registered primitives and components render.
Can we use our own components?
Yes. Leara is designed so your SaaS owns presentation. You register the primitives/components your product supports, then style them through your own design system.
Can we use our own LLM provider?
Yes. The runtime is provider-agnostic. You can bring your own model keys through adapters, while Leara focuses on the interface contract around the model.
What do we pay for?
The SDK is easy to try. Paid plans are for product traffic: generated UI outputs, hosted compose endpoints, traces, usage analytics, limits, and team controls.
Does Leara only work with React?
The core Interface contract is framework-neutral. React is the first renderer because it is the most common path for SaaS apps, but the model/runtime boundary is not React-specific.
When should we use Leara instead of chat?
Use it when the answer should become part of the product: dashboards, reports, tables, comparisons, workflows, forms, and contextual product surfaces. For plain text answers, chat is enough.