Skip to content

The Open Standard for Generative UI

docs: clarify production options and Reliability Monitoring (#1234)

abhithesys3b4151cf95a7
diff --git a/docs/README.md b/docs/README.md
index 0a53ee943..610a6bed4 100644
--- a/docs/README.md
+++ b/docs/README.md
@@ -157,9 +157,10 @@ Active component-preview specific modules are located under `app/components/` an
### Navigation
-The docs use two sidebar modes. The global sidebar introduces the documentation through Start,
+The docs use two sidebar modes. The global sidebar introduces the documentation through Overview,
Build, Production, and Reference groups. Links with chevrons enter a nested section sidebar for
-OpenUI Lang, Build Agents, Agent Interface, Gateway, Observability, or API Reference. Nested page lists continue to come from
+OpenUI Lang, Build Agents (including Agent Interface), Reliability Monitoring, Gateway, or API Reference.
+Production overview and Autofix are direct pages in the global sidebar. Nested page lists continue to come from
each section's `meta.json`; “All docs” restores the global sidebar without changing the current
page. Direct links into a section start in nested mode.
diff --git a/docs/components/gateway-reliability-visualization.tsx b/docs/components/gateway-reliability-visualization.tsx
new file mode 100644
index 000000000..9f9d50cc0
--- /dev/null
+++ b/docs/components/gateway-reliability-visualization.tsx
@@ -0,0 +1,87 @@
+export function GatewayReliabilityVisualization() {
+ return (
+ <div
+ className="not-prose my-6 rounded-2xl border border-fd-border bg-fd-muted/20 p-4 sm:p-6"
+ role="img"
+ aria-label="A request travels from an application through OpenUI Gateway to an LLM. The LLM streams generated UI through OpenUI Gateway, which validates the stream continuously, fixes errors as they are detected, and forwards the validated stream to the application."
+ >
+ <div className="grid items-stretch gap-3 md:grid-cols-[minmax(0,1fr)_7rem_minmax(0,1.15fr)_7rem_minmax(0,1fr)]">
+ <div className="flex min-h-40 flex-col rounded-xl border border-fd-border bg-fd-background p-4">
+ <div className="text-sm font-semibold text-fd-foreground">Your application</div>
+ <div className="mt-4 flex flex-1 flex-col gap-2 rounded-lg border border-fd-border bg-fd-muted/40 p-3">
+ <div className="h-2 w-2/3 rounded-full bg-fd-muted-foreground/25" />
+ <div className="h-2 w-1/2 rounded-full bg-fd-muted-foreground/20" />
+ <div className="mt-2 flex-1 rounded-md border border-[color:color-mix(in_srgb,var(--openui-text-success-primary)_25%,transparent)] bg-[var(--openui-success-background)] p-2">
+ <div className="text-xs font-medium text-[var(--openui-text-success-primary)]">
+ Generated interface
+ </div>
+ <div className="mt-2 h-2 w-full rounded-full bg-fd-muted-foreground/20" />
+ <div className="mt-2 h-2 w-3/4 rounded-full bg-fd-muted-foreground/20" />
+ </div>
+ </div>
+ </div>
+
+ <div className="flex items-center justify-around py-1 text-xs font-medium text-fd-muted-foreground md:hidden">
+ <div className="flex flex-col items-center gap-1">
+ <span>Request</span>
+ <span className="text-lg" aria-hidden="true">
+ ↓
+ </span>
+ </div>
+ <div className="flex flex-col items-center gap-1 text-[var(--openui-text-success-primary)]">
+ <span aria-hidden="true">↑</span>
+ <span>Validated UI</span>
+ </div>
+ </div>
+ <div className="hidden flex-col items-center justify-center gap-10 text-xs font-medium md:flex">
+ <div className="whitespace-nowrap text-fd-muted-foreground">Request →</div>
+ <div className="whitespace-nowrap text-[var(--openui-text-success-primary)]">
+ ← Validated UI
+ </div>
+ </div>
+
+ <div className="flex min-h-40 flex-col rounded-xl border border-[var(--openui-text-purple-primary)] bg-[var(--openui-purple-background)] p-4 shadow-sm">
+ <div className="text-sm font-semibold text-fd-foreground">OpenUI Gateway</div>
+ <div className="text-xs font-medium text-[var(--openui-text-purple-primary)]">
+ Reliability middleware
+ </div>
+ <div className="mt-4 flex flex-1 flex-col justify-center gap-2">
+ <div className="rounded-lg border border-[color:color-mix(in_srgb,var(--openui-text-purple-primary)_30%,transparent)] bg-fd-background/70 px-3 py-2 text-xs font-medium text-fd-foreground">
+ Validate generated UI
+ </div>
+ <div className="rounded-lg border border-[color:color-mix(in_srgb,var(--openui-text-purple-primary)_30%,transparent)] bg-fd-background/70 px-3 py-2 text-xs font-medium text-fd-foreground">
+ Fix errors automatically
+ </div>
+ </div>
+ </div>
+
+ <div className="flex items-center justify-around py-1 text-xs font-medium text-fd-muted-foreground md:hidden">
+ <div className="flex flex-col items-center gap-1">
+ <span>Model request</span>
+ <span className="text-lg" aria-hidden="true">
+ ↓
+ </span>
+ </div>
+ <div className="flex flex-col items-center gap-1 text-[var(--openui-text-alert-primary)]">
+ <span aria-hidden="true">↑</span>
+ <span>Generated UI</span>
+ </div>
+ </div>
+ <div className="hidden flex-col items-center justify-center gap-10 text-xs font-medium text-fd-muted-foreground md:flex">
+ <div className="whitespace-nowrap">Model request →</div>
+ <div className="whitespace-nowrap text-[var(--openui-text-alert-primary)]">
+ ← Generated UI
+ </div>
+ </div>
+
+ <div className="flex min-h-40 flex-col rounded-xl border border-fd-border bg-fd-background p-4">
+ <div className="text-sm font-semibold text-fd-foreground">LLM</div>
+ <div className="text-xs text-fd-muted-foreground">OpenAI, Anthropic, and others</div>
+ <div className="mt-4 flex flex-1 items-center justify-center rounded-lg border border-[color:color-mix(in_srgb,var(--openui-text-alert-primary)_25%,transparent)] bg-[var(--openui-alert-background)] px-3 text-center text-xs font-medium text-[var(--openui-text-alert-primary)]">
+ Generate interface
+ </div>
+ </div>
+ </div>
+ </div>
+ );
+}
diff --git a/docs/content/docs/architecture.mdx b/docs/content/docs/architecture.mdx
index ea67412ae..81fb2a97b 100644
--- a/docs/content/docs/architecture.mdx
+++ b/docs/content/docs/architecture.mdx
@@ -10,13 +10,18 @@ Generative UI fundamentally changes this relationship. Instead of merely providi
## How it works
-<img
- src="/images/openui-architecture.svg"
- alt="A user request travels from an application through OpenUI Gateway to a model provider. The corrected OpenUI Lang stream returns to a renderer, while Observability traces the flow."
- className="not-prose my-8 h-auto w-full overflow-hidden rounded-2xl"
-/>
-
-Your application provides a library of components and the relevant context. OpenUI Gateway sends the request to a model and corrects invalid output. The model uses OpenUI Lang to describe the interface, and the OpenUI Lang Renderer renders it using the provided library. Observability traces this process so you can find failures and improve reliability.
+<figure className="not-prose my-8">
+ <img
+ src="/images/openui-architecture.svg"
+ alt="A user request travels from an application through OpenUI Gateway to a model provider. The corrected OpenUI Lang stream returns to a renderer, with Reliability Monitoring tracking generation failures."
+ className="m-0 block h-auto w-full overflow-hidden rounded-2xl"
+ />
+ <figcaption className="mt-2 text-center text-xs leading-5 text-fd-muted-foreground">
+ Generation through Gateway
+ </figcaption>
+</figure>
+
+Your application provides a library of components and the relevant context. The model generates OpenUI Lang, and the renderer displays it using your components. Gateway can correct output during streaming, or Autofix can repair completed output from your existing inference setup. Reliability Monitoring helps you find failures and improve reliability.
## OpenUI Lang
@@ -129,18 +134,26 @@ OpenUI Gateway provides two OpenAI-compatible APIs:
Learn more in the [OpenUI Gateway documentation](/docs/gateway).
-## Observability
+## Autofix
+
+Reliability for your own inference. Correct model output that would break your interface, without changing your model provider or inference setup.
+
+Learn more in the [Autofix docs](/docs/autofix).
+
+## Reliability Monitoring
Generative UI depends on model providers and model-generated output. Providers can fail, and models can generate invalid or incomplete interfaces.
-OpenUI Observability helps you find these failures in production, trace affected generations, and identify recurring patterns. These insights help you improve your system prompt, component library, and model selection.
+Reliability Monitoring helps you find these failures in production, trace affected generations, and identify recurring patterns. These insights help you improve your system prompt, component library, and model selection.
+
+Monitoring is automatically enabled for Gateway generations and Autofix requests. If you use neither, connect the [Observability SDK](/docs/reliability/installation) to track parsing and rendering errors in your application.
<img
src="/images/openui-lang/reliability-dashboard.png"
- alt="OpenUI Observability dashboard showing generation errors and affected interfaces"
+ alt="Reliability Monitoring dashboard showing generation errors and affected interfaces"
className="w-full rounded-lg border"
/>
-Observability identifies errors but does not correct generated output. Without OpenUI Gateway's validation and correction layer, invalid output can reach users as an incomplete or broken interface.
+Reliability Monitoring identifies errors. Use Gateway or Autofix to correct invalid generated output.
-Learn more in the [OpenUI Observability documentation](/docs/observability).
+Learn more in the [Reliability Monitoring docs](/docs/reliability).
diff --git a/docs/content/docs/gateway/api/autofix.mdx b/docs/content/docs/autofix.mdx
similarity index 95%
rename from docs/content/docs/gateway/api/autofix.mdx
rename to docs/content/docs/autofix.mdx
index e847e27bb..6de2cb8df 100644
--- a/docs/content/docs/gateway/api/autofix.mdx
+++ b/docs/content/docs/autofix.mdx
@@ -3,7 +3,9 @@ title: Autofix API
description: Fix invalid OpenUI Lang from any model with one API call.
---
-The Autofix API repairs invalid OpenUI Lang after a model has generated it. Send the conversation with the generation as the last assistant turn, and the answer carries valid OpenUI Lang. Use it when your application calls a model directly.
+The Autofix API repairs invalid OpenUI Lang after your model has generated it. Keep your existing model provider, inference setup, and agent framework. Send the conversation with the generation as the last assistant turn, and the API returns the repair result. You do not need to generate through Gateway's Chat Completions or Responses API.
+
+Use Autofix for correction after generation, or [Gateway](/docs/gateway) for model access with in-built correction.
**Endpoint:** `POST https://api.thesys.dev/v1/autofix`
@@ -369,7 +371,7 @@ Each entry in `fixed_errors` or `unfixed_errors` names one problem. `code` is al
| `orphaned` | A statement is defined but not reachable from `root`. |
| `missing-root` | The generation has no valid `root` statement. |
-These are the same codes the OpenUI SDK reports in the browser, so an application can handle both with one path. See [Observability](/docs/observability) for how they appear in the Thesys Console.
+These are the same codes the OpenUI SDK reports in the browser, so an application can handle both with one path. See [Reliability Monitoring](/docs/reliability) for how they appear in the Thesys Console.
## Framing
@@ -404,6 +406,6 @@ The endpoint returns the same error shape as the Chat Completions and Responses
| `429` | `rate_limit_error` | The organization has no credits, or billing is suspended. |
| `500` | `internal_server_error` | The fixing model could not be reached. Nothing is charged. |
-## Observability
+## Reliability Monitoring
-Autofix calls appear in the Thesys Console under the model name `openui/autofix`. A successful fix counts as a sanitizer correction on the Reliability page, with the same error codes as corrections made during generation.
+Autofix requests automatically appear in the [Reliability Monitoring](/docs/reliability) under the model name `openui/autofix`. **No Observability SDK** installation is required to track these requests. A successful fix counts as a sanitizer correction, with the same error codes as corrections made during Gateway generation.
diff --git a/docs/content/docs/gateway/index.mdx b/docs/content/docs/gateway/index.mdx
index d3caa147b..94e4ba1bb 100644
--- a/docs/content/docs/gateway/index.mdx
+++ b/docs/content/docs/gateway/index.mdx
@@ -70,9 +70,6 @@ full: true
<Card title="Responses API" href="/docs/gateway/api/responses">
Build stateful model interactions with persistent conversations and hosted tools.
</Card>
- <Card title="Autofix API" href="/docs/gateway/api/autofix">
- Fix invalid OpenUI Lang from any model. Send the raw generation and get back valid OpenUI Lang.
- </Card>
</Cards>
<div style={{maxWidth: '900px'}}>
@@ -81,6 +78,8 @@ full: true
- **Validated generated UI.** Validates generated UI during the stream and fixes malformed syntax, invalid component usage, and schema violations before forwarding the response.
- **Unified API.** Switch between providers and models with minimal code changes.
- **High reliability.** Automatically retries requests with other providers if one fails.
- - **Correction on its own.** Keep your model and backend, and send only the generation that needs fixing to the Autofix API.
+ - **Reliability tracking.** Generations through Gateway automatically appear in the [Reliability Monitoring](/docs/reliability).
- **No markup on tokens.** Tokens cost the same as they would from the provider directly, with zero markup, including with Bring Your Own Key (BYOK).
</div>
+
+Calling a model through your own integration? Use [Autofix](/docs/autofix) to repair completed OpenUI Lang without moving inference to Gateway.
diff --git a/docs/content/docs/gateway/meta.json b/docs/content/docs/gateway/meta.json
index d70318724..4ef188263 100644
--- a/docs/content/docs/gateway/meta.json
+++ b/docs/content/docs/gateway/meta.json
@@ -14,7 +14,6 @@
"---APIs---",
"api/chat-completions",
"api/responses",
- "api/conversations",
- "api/autofix"
+ "api/conversations"
]
}
diff --git a/docs/content/docs/gateway/pricing-credits.mdx b/docs/content/docs/gateway/pricing-credits.mdx
index 021567d9b..4f8e9e3e1 100644
--- a/docs/content/docs/gateway/pricing-credits.mdx
+++ b/docs/content/docs/gateway/pricing-credits.mdx
@@ -40,7 +40,7 @@ Model usage is billed at the provider's rates without markup. The Gateway API an
### Autofix API
-An [Autofix API](/docs/gateway/api/autofix) call that runs a fix costs a flat $0.02, deducted from your Gateway Credits. It does not use a plan API call. A generation that is already valid is free.
+An [Autofix API](/docs/autofix) call that runs a fix costs a flat $0.02, deducted from your Gateway Credits. It does not use a plan API call. A generation that is already valid is free.
### Finding model pricing
diff --git a/docs/content/docs/gateway/reliability.mdx b/docs/content/docs/gateway/reliability.mdx
index 75237d12e..d2d7fadae 100644
--- a/docs/content/docs/gateway/reliability.mdx
+++ b/docs/content/docs/gateway/reliability.mdx
@@ -5,7 +5,7 @@ description: Keep model requests working when providers are unavailable or model
<Callout type="tip" title="New: Autofix API">
Not routing requests through Gateway? Send your model's OpenUI Lang generation to the
- [Autofix API](/docs/gateway/api/autofix) and get back valid OpenUI Lang, at a flat price per
+ [Autofix API](/docs/autofix) and get back valid OpenUI Lang, at a flat price per
fix.
</Callout>
@@ -55,3 +55,7 @@ Gateway improves the reliability of model access and generated OpenUI Lang. Your
- Business rules and data validation
See [OpenUI Lang reliability](/docs/openui-lang/reliability) for component definitions, prompts, evaluations, and renderer safeguards.
+
+## Track reliability
+
+Generations through Gateway automatically appear in the [Reliability Monitoring](/docs/reliability). No Observability SDK installation is required to track those requests. Use the dashboard to inspect affected generations and identify recurring errors.
diff --git a/docs/content/docs/index.mdx b/docs/content/docs/index.mdx
index e1b7c3448..d593dadea 100644
--- a/docs/content/docs/index.mdx
+++ b/docs/content/docs/index.mdx
@@ -10,7 +10,7 @@ OpenUI is a Generative UI toolkit for creating agents that respond with live, in
## OpenUI stack
-OpenUI includes three products for building and running Generative UI:
+OpenUI provides the building blocks for creating generated interfaces and running them reliably in production.
<Cards>
<Card title="OpenUI Lang" href="/docs/openui-lang">
@@ -19,13 +19,18 @@ OpenUI includes three products for building and running Generative UI:
tokens**](/docs/openui-lang/benchmarks) than JSON.
</Card>
<Card title="Gateway" href="/docs/gateway">
- OpenRouter for Generative UI. Use one API to access models across providers, with automatic
- failover and model-output correction that prevents broken interfaces.
+ OpenRouter for Generative UI. Access models across providers with automatic fallbacks and
+ correction of invalid OpenUI Lang during streaming.
</Card>
- <Card title="Observability" href="/docs/observability">
- Production visibility for Generative UI. Trace generated interfaces and improve their
- reliability.
+ <Card title="Autofix" href="/docs/autofix">
+ Reliability for your own inference. Correct model output that would break your interface,
+ without changing your model provider or inference setup.
+ </Card>
+ <Card title="Reliability Monitoring" href="/docs/reliability">
+ Track generation failures, investigate affected interfaces, and identify recurring errors.
</Card>
</Cards>
+Learn about correction and monitoring options in [OpenUI in production](/docs/production).
+
Ready to build? [Get started with OpenUI](/docs/getting-started).
diff --git a/docs/content/docs/meta.json b/docs/content/docs/meta.json
index e6a248f5c..eb73a0e94 100644
--- a/docs/content/docs/meta.json
+++ b/docs/content/docs/meta.json
@@ -7,8 +7,10 @@
"openui-lang",
"build-agents",
"agent",
+ "production",
"gateway",
- "observability",
+ "autofix",
+ "reliability",
"deploy",
"api-reference"
]
diff --git a/docs/content/docs/observability/index.mdx b/docs/content/docs/observability/index.mdx
deleted file mode 100644
index c85ea1e22..000000000
--- a/docs/content/docs/observability/index.mdx
+++ /dev/null
@@ -1,12 +0,0 @@
----
-title: Overview
-description: Find and understand OpenUI generation and rendering errors across production traffic.
----
-
-Generative UI turns model output into interfaces that people use directly. Models usually follow the component contract they receive, but they can still invent components, use unsupported values, leave references unresolved, or stop early. This can produce incomplete or broken interfaces.
-
-![OpenUI reliability dashboard](/images/observability/reliability-dashboard.png)
-
-In our [**Reliability benchmark**](/blog/generative-ui-benchmark), every tested format, including OpenUI, A2UI, and json-render, produced incomplete generations. OpenUI completed 92.9% of evaluated tasks and produced 2 blank screens across 1,104 runs. Its strongest model pairing reached 99.5%, not 100%.
-
-OpenUI Observability helps teams find and trace these failures after deployment so they can improve the reliability of generated interfaces.
diff --git a/docs/content/docs/observability/installation.mdx b/docs/content/docs/observability/installation.mdx
deleted file mode 100644
index 72118588d..000000000
--- a/docs/content/docs/observability/installation.mdx
+++ /dev/null
@@ -1,48 +0,0 @@
----
-title: Installation
-description: Connect an OpenUI application to hosted Observability with the official production SDK.
----
-
-Install `@openuidev/observability-cloud` in the application that renders OpenUI interfaces. The SDK sends supported runtime telemetry to the hosted Observability product.
-
-## 1. Install the SDK
-
-```bash tab="pnpm" tab-group="pkg"
-pnpm add @openuidev/observability-cloud
-```
-
-```bash tab="bun" tab-group="pkg"
-bun add @openuidev/observability-cloud
-```
-
-```bash tab="yarn" tab-group="pkg"
-yarn add @openuidev/observability-cloud
-```
-
-```bash tab="npm" tab-group="pkg"
-npm install @openuidev/observability-cloud
-```
-
-## 2. Create a client key
-
-Create a client API key in the [Thesys Console](https://console.thesys.dev/client-api-keys). Client keys are intended for application runtime telemetry; do not substitute a Gateway server API key.
-
-## 3. Initialize once
-
-Initialize Observability at the root of the application before rendering OpenUI responses:
-
-```ts
-import * as Observability from "@openuidev/observability-cloud";
-
-Observability.init({
- apiKey: "pk-th-…",
-});
-```
-
-Deploy the application and exercise representative generative UI flows. Runtime errors will begin appearing in the Console's reliability views.
-
-## Works with any generation path
-
-The SDK captures supported OpenUI runtime signals whether the application generated the response through the [Gateway](/docs/gateway) or called a model provider directly.
-
-Continue with [Dashboard](/docs/observability/dashboard) to interpret the data.
diff --git a/docs/content/docs/openui-lang/overview.mdx b/docs/content/docs/openui-lang/overview.mdx
index 46c830b48..7e7817f25 100644
--- a/docs/content/docs/openui-lang/overview.mdx
+++ b/docs/content/docs/openui-lang/overview.mdx
@@ -21,13 +21,13 @@ Four pieces make that flow work:
- **Renderer:** The renderer maps the parsed tree to your components so the result uses your design
system and application behavior.
-<GatewayCallout description="Use the Gateway as the default production generation path, then add Observability to inspect errors from real traffic." />
+<GatewayCallout description="For production, route generation through the Gateway to correct invalid OpenUI Lang while it streams and recover from provider failures." />
## The core loop
1. Define the components your application supports.
2. Generate model instructions from those definitions.
-3. Send the instructions and conversation through the Gateway.
+3. Send the instructions and conversation to a model provider or route through [Gateway](/docs/gateway) for reliability.
4. Stream the resulting OpenUI Lang into the renderer.
5. Handle component interactions with application state, tools, queries, and mutations.
diff --git a/docs/content/docs/openui-lang/reliability.mdx b/docs/content/docs/openui-lang/reliability.mdx
index 3372b6c93..68e1696eb 100644
--- a/docs/content/docs/openui-lang/reliability.mdx
+++ b/docs/content/docs/openui-lang/reliability.mdx
@@ -3,6 +3,8 @@ title: Reliability
description: Build reliable LLM-generated interfaces by measuring, reducing, and correcting generation errors.
---
+import { GatewayReliabilityVisualization } from "@/components/gateway-reliability-visualization";
+
Generative UI turns LLM output directly into interfaces that users see and interact with. The model receives a description of the available components and their properties, then generates an interface that follows that contract. Most generations follow it correctly, but an LLM can occasionally invent a component, use an unsupported value, leave a reference unresolved, or stop before completing the interface.
<img
@@ -17,7 +19,7 @@ In our [**Reliability benchmark**](/blog/generative-ui-benchmark), every tested
<img
src="/images/openui-lang/reliability-layers.svg"
- alt="Reliability pipeline showing schema and prompt improvements, model selection, validation and correction, and observability feedback"
+ alt="Reliability pipeline showing schema and prompt improvements, model selection, validation and correction, and reliability monitoring feedback"
className="w-full"
/>
@@ -26,7 +28,7 @@ There are four main ways to improve the reliability of LLM-generated interfaces:
- **Simplify the component schema.** Use distinct component names, clear descriptions, focused properties, and unambiguous enum values. Remove overlapping components and group related ones with `componentGroups`. See [Defining Components](/docs/openui-lang/defining-components).
- **Refine the system prompt.** Add targeted rules for recurring errors and provide valid examples for combinations the model finds difficult. In our benchmark, a single rule improved one model's completion rate by 13 percentage points, while an incorrect example caused double-digit regressions. See [System Prompts](/docs/openui-lang/system-prompts).
- **Evaluate model selection.** Test models using your own component library and representative prompts. Run each prompt multiple times and compare reliability alongside latency and cost.
-- **Validate and correct generated output.** Detect invalid output at runtime and fix it before it reaches users. OpenUI Gateway provides this reliability layer for production applications.
+- **Validate and correct generated output.** We recommend [Gateway](/docs/gateway) for correction during streaming, provider fallbacks, and automatic Reliability Monitoring. If you keep your existing inference setup, use [Autofix](/docs/autofix) to fix invalid generations.
Measure reliability before and after making these changes so you can identify the right intervention and verify that it worked.
@@ -55,60 +57,60 @@ Learn more in [developer tools docs](/docs/openui-lang/developer-tools).
### In production
-Development testing cannot cover every prompt your users will submit. Use `@openuidev/observability-cloud` to monitor UI generation across production traffic and review errors in the Thesys Console.
+Use [Reliability Monitoring](/docs/reliability) to track generation failures, identify recurring errors, and inspect the affected generations. Follow the [setup guide](/docs/reliability/installation) to enable monitoring for your application.
<img
src="/images/openui-lang/reliability-dashboard.png"
- alt="Reliability pipeline showing schema and prompt improvements, model selection, validation and correction, and observability feedback"
+ alt="Reliability Monitoring dashboard showing generation errors and affected interfaces"
className="w-full rounded-lg border"
/>
-Install the package:
+Use these findings to prioritize improvements to your component library, system prompt, and model selection.
-```bash tab="pnpm" tab-group="pkg"
-pnpm install @openuidev/observability-cloud
-```
+<Callout type="tip">
+ **Using a coding agent?** The OpenUI skill can help your agent configure Reliability Monitoring.
+ [Install the OpenUI skill](/docs/mcp#install-via-the-skills-cli-recommended).
+</Callout>
-```bash tab="bun" tab-group="pkg"
-bun install @openuidev/observability-cloud
-```
+## Production reliability
-```bash tab="yarn" tab-group="pkg"
-yarn install @openuidev/observability-cloud
-```
+Even with a well-tuned schema, prompt, and model, some generations in production will still fail validation.
-```bash tab="npm" tab-group="pkg"
-npm run install @openuidev/observability-cloud
-```
+- **[OpenUI Gateway](#openui-gateway) (recommended):** correct output during streaming, recover from provider failures, and monitor generation reliability through one integration.
+- **[Autofix API](#autofix-api):** keep your existing inference setup and send completed generations for repair. You can start by monitoring errors with the Observability SDK, but SDK instrumentation is not required to use Autofix.
-Create a client API key from the [Thesys Console](https://console.thesys.dev/client-api-keys), then initialize observability once in your application:
+### OpenUI Gateway
-```ts
-import * as Observability from "@openuidev/observability-cloud";
+OpenUI Gateway sits between your application and the LLM. It validates generated UI and fixes errors as it streams to your application.
-Observability.init({
- apiKey: "pk-th-…",
-});
-```
+<GatewayReliabilityVisualization />
+
+For every generation, OpenUI Gateway:
-After deploying, open the [Reliability dashboard](https://console.thesys.dev/reliability) to see how frequently generation errors occur, identify the most common error types, and inspect the affected UI generations. This helps you prioritize the schema, prompt, or model changes that will have the greatest impact.
+- **Validates** generated UI as it streams.
+- **Fixes** malformed output and schema violations on the fly.
+- **Recovers** from model or provider failures using fallbacks.
+- **Tracks** generation reliability automatically in [Reliability Monitoring](/docs/reliability).
+
+In production data from our [Reliability benchmark](/blog/generative-ui-benchmark), OpenUI Gateway fixed **96% of generations** that initially failed validation.
<Callout type="tip">
- **Using a coding agent?** The OpenUI skill can help your agent add production observability and
- configure reliability monitoring. [Install the OpenUI
- skill](/docs/mcp#install-via-the-skills-cli-recommended).
+ **Using a coding agent?** The OpenUI skill can help your agent integrate OpenUI Gateway into your
+ application. [Install the OpenUI skill](/docs/mcp#install-via-the-skills-cli-recommended).
</Callout>
-## Production reliability
+#### Integrations options
-Even with a well-tuned schema, prompt, and model, some generations in production will still fail validation. OpenUI offers two ways to repair them automatically, depending on how your application calls the model.
+OpenUI Gateway provides two OpenAI-compatible APIs for this generation flow:
+
+- **[Chat Completions](/docs/gateway/api/chat-completions):** A drop-in endpoint for applications that already use the OpenAI Chat Completions API for generative UI.
+- **[Responses API](/docs/gateway/api/responses):** Build stateful model interactions that continue from previous responses, with built-in persistence and hosted tools such as web search, image search, presentation generation, and more.
-- **[Autofix API](#autofix-api):** your application calls the model itself and sends the generation to the Autofix API for repair.
-- **[OpenUI Gateway](#openui-gateway):** OpenUI Gateway calls the model on your behalf and repairs the generation while it streams.
+Learn more in the [OpenUI Gateway documentation](/docs/gateway).
### Autofix API
-The [Autofix API](/docs/gateway/api/autofix) repairs OpenUI Lang after your model has generated it, so your application keeps its own model calls, keys, and prompts.
+The [Autofix API](/docs/autofix) repairs OpenUI Lang after your model has generated it, so your application keeps its own model calls, keys, and prompts.
For every generation, the Autofix API:
@@ -144,111 +146,4 @@ const completion = await client.chat.completions.create({
`openui.spec.json` is the component-library spec generated by `openui generate --spec`. Autofix reads it from the first system or developer message; do not pass a top-level `library` field. See [Generate OpenUI Lang](/docs/gateway/generate-openui-lang) for how to generate the spec.
-A generation that is already valid is free, and each fix is charged at a flat price. See the [Autofix API reference](/docs/gateway/api/autofix) for the full request and response format and the error codes.
-
-### OpenUI Gateway
-
-OpenUI Gateway sits between your application and the LLM. It validates generated UI and fixes errors as it streams to your application.
-
-<div
- className="not-prose my-6 rounded-2xl border border-fd-border bg-fd-muted/20 p-4 sm:p-6"
- role="img"
- aria-label="A request travels from an application through OpenUI Gateway to an LLM. The LLM streams generated UI through OpenUI Gateway, which validates the stream continuously, fixes errors as they are detected, and forwards the validated stream to the application."
->
- <div className="grid items-stretch gap-3 md:grid-cols-[minmax(0,1fr)_7rem_minmax(0,1.15fr)_7rem_minmax(0,1fr)]">
- <div className="flex min-h-40 flex-col rounded-xl border border-fd-border bg-fd-background p-4">
- <div className="text-sm font-semibold text-fd-foreground">Your application</div>
- <div className="mt-4 flex flex-1 flex-col gap-2 rounded-lg border border-fd-border bg-fd-muted/40 p-3">
- <div className="h-2 w-2/3 rounded-full bg-fd-muted-foreground/25" />
- <div className="h-2 w-1/2 rounded-full bg-fd-muted-foreground/20" />
- <div className="mt-2 flex-1 rounded-md border border-[color:color-mix(in_srgb,var(--openui-text-success-primary)_25%,transparent)] bg-[var(--openui-success-background)] p-2">
- <div className="text-xs font-medium text-[var(--openui-text-success-primary)]">
- Generated interface
- </div>
- <div className="mt-2 h-2 w-full rounded-full bg-fd-muted-foreground/20" />
- <div className="mt-2 h-2 w-3/4 rounded-full bg-fd-muted-foreground/20" />
- </div>
- </div>
- </div>
-
- <div className="flex items-center justify-around py-1 text-xs font-medium text-fd-muted-foreground md:hidden">
- <div className="flex flex-col items-center gap-1">
- <span>Request</span>
- <span className="text-lg" aria-hidden="true">↓</span>
- </div>
- <div className="flex flex-col items-center gap-1 text-[var(--openui-text-success-primary)]">
- <span aria-hidden="true">↑</span>
- <span>Validated UI</span>
- </div>
- </div>
- <div className="hidden flex-col items-center justify-center gap-10 text-xs font-medium md:flex">
- <div className="whitespace-nowrap text-fd-muted-foreground">Request →</div>
- <div className="whitespace-nowrap text-[var(--openui-text-success-primary)]">
- ← Validated UI
- </div>
- </div>
-
- <div className="flex min-h-40 flex-col rounded-xl border border-[var(--openui-text-purple-primary)] bg-[var(--openui-purple-background)] p-4 shadow-sm">
- <div className="text-sm font-semibold text-fd-foreground">OpenUI Gateway</div>
- <div className="text-xs font-medium text-[var(--openui-text-purple-primary)]">
- Reliability middleware
- </div>
- <div className="mt-4 flex flex-1 flex-col justify-center gap-2">
- <div className="rounded-lg border border-[color:color-mix(in_srgb,var(--openui-text-purple-primary)_30%,transparent)] bg-fd-background/70 px-3 py-2 text-xs font-medium text-fd-foreground">
- Validate generated UI
- </div>
- <div className="rounded-lg border border-[color:color-mix(in_srgb,var(--openui-text-purple-primary)_30%,transparent)] bg-fd-background/70 px-3 py-2 text-xs font-medium text-fd-foreground">
- Fix errors automatically
- </div>
- </div>
- </div>
-
- <div className="flex items-center justify-around py-1 text-xs font-medium text-fd-muted-foreground md:hidden">
- <div className="flex flex-col items-center gap-1">
- <span>Model request</span>
- <span className="text-lg" aria-hidden="true">↓</span>
- </div>
- <div className="flex flex-col items-center gap-1 text-[var(--openui-text-alert-primary)]">
- <span aria-hidden="true">↑</span>
- <span>Generated UI</span>
- </div>
- </div>
- <div className="hidden flex-col items-center justify-center gap-10 text-xs font-medium text-fd-muted-foreground md:flex">
- <div className="whitespace-nowrap">Model request →</div>
- <div className="whitespace-nowrap text-[var(--openui-text-alert-primary)]">
- ← Generated UI
- </div>
- </div>
-
- <div className="flex min-h-40 flex-col rounded-xl border border-fd-border bg-fd-background p-4">
- <div className="text-sm font-semibold text-fd-foreground">LLM</div>
- <div className="text-xs text-fd-muted-foreground">OpenAI, Anthropic, and others</div>
- <div className="mt-4 flex flex-1 items-center justify-center rounded-lg border border-[color:color-mix(in_srgb,var(--openui-text-alert-primary)_25%,transparent)] bg-[var(--openui-alert-background)] px-3 text-center text-xs font-medium text-[var(--openui-text-alert-primary)]">
- Generate interface
- </div>
- </div>
-
- </div>
-</div>
-
-For every generation, OpenUI Gateway:
-
-- **Validates** generated UI as it streams.
-- **Fixes** malformed output and schema violations on the fly.
-- **Recovers** from model or provider failures using fallbacks.
-
-In production data from our [Reliability benchmark](/blog/generative-ui-benchmark), OpenUI Gateway fixed **96% of generations** that initially failed validation.
-
-<Callout type="tip">
- **Using a coding agent?** The OpenUI skill can help your agent integrate OpenUI Gateway into your
- application. [Install the OpenUI skill](/docs/mcp#install-via-the-skills-cli-recommended).
-</Callout>
-
-#### Integrations options
-
-OpenUI Gateway provides two OpenAI-compatible APIs for this generation flow:
-
-- **[Chat Completions](/docs/gateway/api/chat-completions):** A drop-in endpoint for applications that already use the OpenAI Chat Completions API for generative UI.
-- **[Responses API](/docs/gateway/api/responses):** Build stateful model interactions that continue from previous responses, with built-in persistence and hosted tools such as web search, image search, presentation generation, and more.
-
-Learn more in the [OpenUI Gateway documentation](/docs/gateway).
+A generation that is already valid is free, and each fix is charged at a flat price. See the [Autofix API reference](/docs/autofix) for the full request and response format and the error codes.
diff --git a/docs/content/docs/openui-lang/system-prompts.mdx b/docs/content/docs/openui-lang/system-prompts.mdx
index ac4a5e449..87794fcd8 100644
--- a/docs/content/docs/openui-lang/system-prompts.mdx
+++ b/docs/content/docs/openui-lang/system-prompts.mdx
@@ -85,9 +85,11 @@ const systemPrompt = generateSystemPrompt({
| `editMode` | Incremental editing - LLM outputs only changed statements | `false` |
| `inlineMode` | Text + fenced code responses - LLM can answer questions without generating UI | `false` |
-### OpenUI Gateway
+### Gateway and Autofix
-Pass `cloud: true` with the generated library specification to emit Gateway's managed config block instead of a local prompt. The option name is retained for compatibility. OpenUI Gateway assembles the full system prompt on the server, and `instructions` adds extra prose after the config block.
+Pass `cloud: true` with the generated library specification to emit a configuration block for Gateway or Autofix. Gateway uses it to assemble the system prompt; Autofix uses it to validate and repair generated OpenUI Lang against your component library. The option name is retained for compatibility.
+
+For Gateway requests, `instructions` adds extra prose after the configuration block.
```ts
import { generateSystemPrompt } from "@openuidev/lang-core";
@@ -109,7 +111,7 @@ const instructions = generateSystemPrompt({
});
```
-See the [Generative UI Quickstart](/docs/gateway/generate-openui-lang) for the Gateway workflow.
+See the [Generative UI Quickstart](/docs/gateway/generate-openui-lang) for the Gateway workflow, or the [Autofix API](/docs/autofix) to repair output from your existing model integration.
Built-in functions (`@Count`, `@Filter`, `@Sort`, `@Each`, etc.) are automatically included when either `toolCalls` or `bindings` is enabled. For static UI libraries without data fetching, they are omitted to keep the prompt concise.
diff --git a/docs/content/docs/production.mdx b/docs/content/docs/production.mdx
new file mode 100644
index 000000000..2908e2e24
--- /dev/null
+++ b/docs/content/docs/production.mdx
@@ -0,0 +1,85 @@
+---
+title: OpenUI in production
+description: Track generation failures, fix invalid OpenUI Lang, and keep model requests working in production.
+---
+
+import { GatewayReliabilityVisualization } from "@/components/gateway-reliability-visualization";
+
+Running Generative UI reliably in production takes more than getting an interface to render once. Models can use unknown components, supply invalid properties, leave references unresolved, or stop before completing a response. Model Provider outages can also interrupt generation.
+
+In our [**reliability benchmark**](/blog/generative-ui-benchmark), OpenUI achieved the highest overall structural validity and the fewest blank screens among the tested frameworks, but no framework was error-free.
+
+Production applications need ways to track these failures, fix invalid output, and recover from provider interruptions. OpenUI provides the following tools:
+
+<Cards>
+ <Card title="Gateway" href="/docs/gateway">
+ OpenRouter for Generative UI. Access models across providers with automatic fallbacks and correction of invalid OpenUI Lang during streaming.
+ </Card>
+ <Card title="Autofix" href="/docs/autofix">
+ Fix invalid OpenUI Lang while keeping your existing model provider and inference setup.
+ </Card>
+ <Card title="Reliability Monitoring" href="/docs/reliability">
+ Track generation failures, investigate affected interfaces, and identify recurring errors.
+ </Card>
+</Cards>
+
+## Gateway
+
+OpenUI Gateway provides a unified API to access models across providers through a single endpoint. It sits between your application and the LLM, validates generated OpenUI Lang, and fixes errors as the response streams to your application. It also provides automatic provider fallbacks.
+
+<GatewayReliabilityVisualization />
+
+Integrate OpenUI Gateway with your existing backend stack, including LangGraph, Vercel AI SDK, or the OpenAI SDK. Gateway provides OpenAI-compatible Chat Completions and Responses APIs:
+
+<Tabs items={["Chat Completions", "Responses API"]}>
+ <Tab value="Chat Completions">
+ ```ts title="server.ts"
+ import OpenAI from "openai";
+
+ const gateway = new OpenAI({
+ apiKey: process.env.THESYS_API_KEY,
+ baseURL: "https://api.thesys.dev/v1/embed",
+ });
+
+ const completion = await gateway.chat.completions.create({
+ model: "openai/gpt-5",
+ messages: [
+ { role: "user", content: "Compare quarterly revenue by region." },
+ ],
+ stream: true,
+ });
+ ```
+ </Tab>
+ <Tab value="Responses API">
+ ```ts title="server.ts"
+ import OpenAI from "openai";
+
+ const gateway = new OpenAI({
+ apiKey: process.env.THESYS_API_KEY,
+ baseURL: "https://api.thesys.dev/v1/embed",
+ });
+
+ const response = await gateway.responses.create({
+ model: "openai/gpt-5",
+ input: "Compare quarterly revenue by region.",
+ stream: true,
+ });
+ ```
+ </Tab>
+</Tabs>
+
+Learn more about OpenUI Gateway [here](/docs/gateway).
+
+## Autofix
+
+When you call a model provider directly, invalid OpenUI Lang reaches your application without any correction. The Autofix API fixes it while you keep your existing inference setup. After generation completes, validate the output with the OpenUI Lang parser. If validation fails, send the generation, component specification, and conversation context to Autofix, which returns corrected OpenUI Lang.
+
+Learn more in the [Autofix docs](/docs/autofix).
+
+## Reliability Monitoring
+
+Fixing errors keeps interfaces working, but it doesn't tell you why they fail. Reliability Monitoring tracks how often generated interfaces fail and which errors recur, with insights into how to fix them. Inspect the affected OpenUI Lang to find the root cause, then improve your component library, system prompt, or model selection.
+
+Monitoring is automatically enabled for Gateway generations and Autofix requests. If you use neither, connect the Observability SDK to track parsing and rendering errors in your application.
+
+Learn more in the [Reliability Monitoring docs](/docs/reliability).
diff --git a/docs/content/docs/observability/dashboard.mdx b/docs/content/docs/reliability/dashboard.mdx
similarity index 80%
rename from docs/content/docs/observability/dashboard.mdx
rename to docs/content/docs/reliability/dashboard.mdx
index 7637415db..bd0aa43a4 100644
--- a/docs/content/docs/observability/dashboard.mdx
+++ b/docs/content/docs/reliability/dashboard.mdx
@@ -1,9 +1,9 @@
---
title: Dashboard
-description: Use the hosted dashboard to find recurring OpenUI errors and inspect affected generations.
+description: Find recurring OpenUI errors and inspect affected generations in Reliability Monitoring.
---
-After production setup, open the [Reliability dashboard](https://console.thesys.dev/reliability) to understand how generated interfaces behave across real requests.
+Open the [Reliability dashboard](https://console.thesys.dev/reliability) to understand how generated interfaces behave across real requests. Tracking is automatic for Gateway generations and Autofix requests. For other generation paths, follow [Setup](/docs/reliability/installation) to connect the Observability SDK.
Use the dashboard to answer three questions:
@@ -35,4 +35,4 @@ Click on an error occurance to inspect the errors associated with a specific gen
The detail view shows when the occurrence was recorded, and the generated OpenUI Lang that caused the failure. Each error includes a code, a readable explanation, and relevant details such as the component or statement involved.
-See [Optimizing with Observability](/docs/observability/optimizing) for a repeatable improvement loop.
+See [Improve reliability](/docs/reliability/optimizing) for a repeatable improvement loop.
diff --git a/docs/content/docs/reliability/index.mdx b/docs/content/docs/reliability/index.mdx
new file mode 100644
index 000000000..33c8d69b6
--- /dev/null
+++ b/docs/content/docs/reliability/index.mdx
@@ -0,0 +1,24 @@
+---
+title: Reliability Monitoring
+description: Find and understand OpenUI generation and rendering errors across production traffic.
+---
+
+Reliability Monitoring shows how often generated interfaces fail, which errors recur, and the generation associated with each failure, with insights into how to fix them. Use it to decide which components, prompt rules, or model choices need attention.
+
+![OpenUI Reliability Monitoring dashboard](/images/observability/reliability-dashboard.png)
+
+## Enable monitoring
+
+How you connect depends on how you generate OpenUI Lang:
+
+| Generation path | Setup |
+| --------------- | ----- |
+| [Gateway](/docs/gateway) or [Autofix](/docs/autofix) | Monitoring is enabled automatically for generations and requests through these services. |
+| Your own inference without Autofix | Install the [Observability SDK](/docs/reliability/installation) to report errors from your application. |
+
+## Investigate failures
+
+Open the [Reliability dashboard](https://console.thesys.dev/reliability) to review errors and inspect the generation associated with each occurrence.
+
+- The [dashboard guide](/docs/reliability/dashboard) explains each view and the error details.
+- [Improve reliability](/docs/reliability/optimizing) shows how to turn recurring failures into changes to your components, prompts, and model selection.
diff --git a/docs/content/docs/reliability/installation.mdx b/docs/content/docs/reliability/installation.mdx
new file mode 100644
index 000000000..cd313405e
--- /dev/null
+++ b/docs/content/docs/reliability/installation.mdx
@@ -0,0 +1,52 @@
+---
+title: Setup
+description: Enable reliability tracking through Gateway, Autofix, or the Observability SDK.
+---
+
+## Using Gateway or Autofix
+
+Reliability tracking is automatically enabled for generations through [Gateway](/docs/gateway) and requests sent to [Autofix](/docs/autofix). Open the [Reliability dashboard](https://console.thesys.dev/reliability) to inspect those requests. No SDK installation is required for this tracking.
+
+## Using your own model integration
+
+If you use neither Gateway nor Autofix, install `@openuidev/observability-cloud` in the application that renders OpenUI interfaces. The Observability SDK sends supported runtime telemetry to Reliability Monitoring.
+
+You can also use the SDK alongside Gateway or Autofix to capture supported parsing and rendering signals from your application.
+
+### 1. Install the SDK
+
+```bash tab="pnpm" tab-group="pkg"
+pnpm add @openuidev/observability-cloud
+```
+
+```bash tab="bun" tab-group="pkg"
+bun add @openuidev/observability-cloud
+```
+
+```bash tab="yarn" tab-group="pkg"
+yarn add @openuidev/observability-cloud
+```
+
+```bash tab="npm" tab-group="pkg"
+npm install @openuidev/observability-cloud
+```
+
+### 2. Create a client key
+
+Create a client API key in the [Thesys Console](https://console.thesys.dev/client-api-keys). Client keys are intended for application runtime telemetry; do not substitute a Gateway server API key.
+
+### 3. Initialize once
+
+Initialize the SDK at the root of the application before rendering OpenUI responses:
+
+```ts
+import * as Observability from "@openuidev/observability-cloud";
+
+Observability.init({
+ apiKey: "pk-th-…",
+});
+```
+
+Deploy the application and exercise representative generative UI flows. Runtime errors will begin appearing in the Console's reliability views.
+
+Continue with [Dashboard](/docs/reliability/dashboard) to interpret the data.
diff --git a/docs/content/docs/observability/meta.json b/docs/content/docs/reliability/meta.json
similarity index 53%
rename from docs/content/docs/observability/meta.json
rename to docs/content/docs/reliability/meta.json
index 10b2741d6..e1632affe 100644
--- a/docs/content/docs/observability/meta.json
+++ b/docs/content/docs/reliability/meta.json
@@ -1,8 +1,8 @@
{
- "title": "Observability",
+ "title": "Reliability Monitoring",
"root": true,
"pages": [
- "index",
+ "[Introduction](/docs/reliability)",
"installation",
"dashboard",
"optimizing"
diff --git a/docs/content/docs/observability/optimizing.mdx b/docs/content/docs/reliability/optimizing.mdx
similarity index 88%
rename from docs/content/docs/observability/optimizing.mdx
rename to docs/content/docs/reliability/optimizing.mdx
index e3dfe5f24..a851ac4f3 100644
--- a/docs/content/docs/observability/optimizing.mdx
+++ b/docs/content/docs/reliability/optimizing.mdx
@@ -1,9 +1,9 @@
---
-title: Optimizing with Observability
+title: Improve reliability
description: Turn recurring runtime errors into focused improvements to components, prompts, and model selection.
---
-Observability is most useful as a feedback loop, not only an incident log. Use production evidence to choose the smallest change likely to improve reliability, then verify the result against the same signal.
+Use Reliability Monitoring as a feedback loop. Choose a focused change based on recurring production errors, then verify the result against the same signal.
## 1. Find a repeated failure
diff --git a/docs/lib/docs-navigation.test.ts b/docs/lib/docs-navigation.test.ts
index 8b959c327..c8569f37a 100644
--- a/docs/lib/docs-navigation.test.ts
+++ b/docs/lib/docs-navigation.test.ts
@@ -56,11 +56,13 @@ describe("global docs navigation", () => {
children: undefined,
},
{ type: "separator", name: "Production", url: undefined, children: undefined },
+ { type: "page", name: "Overview", url: "/docs/production", children: undefined },
{ type: "page", name: "Gateway", url: "/docs/gateway", children: undefined },
+ { type: "page", name: "Autofix", url: "/docs/autofix", children: undefined },
{
type: "page",
- name: "Observability",
- url: "/docs/observability",
+ name: "Reliability Monitoring",
+ url: "/docs/reliability",
children: undefined,
},
{ type: "page", name: "Deploy your app", url: "/docs/deploy", children: undefined },
@@ -79,13 +81,17 @@ describe("global docs navigation", () => {
assert.equal(getNestedRootForEntryUrl("/docs/build-agents"), "build-agents");
assert.equal(getNestedRootForEntryUrl("/docs/agent/getting-started/introduction"), undefined);
assert.equal(getNestedRootForEntryUrl("/docs/gateway"), "gateway");
- assert.equal(getNestedRootForEntryUrl("/docs/observability"), "observability");
+ assert.equal(getNestedRootForEntryUrl("/docs/autofix"), undefined);
+ assert.equal(getNestedRootForEntryUrl("/docs/production"), undefined);
+ assert.equal(getNestedRootForEntryUrl("/docs/reliability"), "reliability");
assert.equal(getNestedRootForEntryUrl("/docs/api-reference"), "api-reference");
assert.equal(getNestedRootForEntryUrl("/docs/openui-lang/quickstart"), undefined);
});
it("uses a nested sidebar for direct links into a nested section", () => {
assert.deepEqual(getDefaultSidebarMode("/docs"), { kind: "global" });
+ assert.deepEqual(getDefaultSidebarMode("/docs/autofix"), { kind: "global" });
+ assert.deepEqual(getDefaultSidebarMode("/docs/production"), { kind: "global" });
assert.deepEqual(getDefaultSidebarMode("/docs/overview"), { kind: "global" });
assert.deepEqual(getDefaultSidebarMode("/docs/getting-started"), { kind: "global" });
assert.deepEqual(getDefaultSidebarMode("/docs/openui-lang/comparison"), { kind: "global" });
@@ -113,9 +119,9 @@ describe("global docs navigation", () => {
kind: "nested",
root: "gateway",
});
- assert.deepEqual(getDefaultSidebarMode("/docs/observability/installation"), {
+ assert.deepEqual(getDefaultSidebarMode("/docs/reliability/installation"), {
kind: "nested",
- root: "observability",
+ root: "reliability",
});
assert.deepEqual(getDefaultSidebarMode("/docs/mcp"), { kind: "global" });
});
@@ -149,6 +155,8 @@ describe("global docs navigation", () => {
it("promotes overview pages while grouping product pages under their roots", () => {
assert.equal(getGlobalActiveItemUrl("/docs/getting-started"), "/docs/getting-started");
+ assert.equal(getGlobalActiveItemUrl("/docs/autofix"), "/docs/autofix");
+ assert.equal(getGlobalActiveItemUrl("/docs/production"), "/docs/production");
assert.equal(
getGlobalActiveItemUrl("/docs/openui-lang/comparison"),
"/docs/openui-lang/comparison",
@@ -161,7 +169,7 @@ describe("global docs navigation", () => {
);
assert.equal(getGlobalActiveItemUrl("/docs/build-agents/copilotkit"), "/docs/build-agents");
assert.equal(getGlobalActiveItemUrl("/docs/gateway/api/responses"), "/docs/gateway");
- assert.equal(getGlobalActiveItemUrl("/docs/observability/dashboard"), "/docs/observability");
+ assert.equal(getGlobalActiveItemUrl("/docs/reliability/dashboard"), "/docs/reliability");
});
});
diff --git a/docs/lib/docs-navigation.ts b/docs/lib/docs-navigation.ts
index 3b84183d3..1ca5f9416 100644
--- a/docs/lib/docs-navigation.ts
+++ b/docs/lib/docs-navigation.ts
@@ -1,6 +1,6 @@
import type * as PageTree from "fumadocs-core/page-tree";
-export type NestedDocsRoot = "openui-lang" | "build-agents" | "gateway" | "observability" | "api-reference";
+export type NestedDocsRoot = "openui-lang" | "build-agents" | "gateway" | "reliability" | "api-reference";
export type SidebarMode =
| { kind: "global" }
@@ -40,11 +40,11 @@ export const NESTED_DOCS_SECTIONS: Record<NestedDocsRoot, NestedSection> = {
pathPrefix: "/docs/gateway",
treeFolder: "gateway",
},
- observability: {
- title: "Observability",
- entryUrl: "/docs/observability",
- pathPrefix: "/docs/observability",
- treeFolder: "observability",
+ reliability: {
+ title: "Reliability Monitoring",
+ entryUrl: "/docs/reliability",
+ pathPrefix: "/docs/reliability",
+ treeFolder: "reliability",
},
"api-reference": {
title: "API Reference",
@@ -61,6 +61,8 @@ const promotedGlobalUrls = new Set([
"/docs/openui-lang/comparison",
"/docs/mcp",
"/docs/deploy",
+ "/docs/production",
+ "/docs/autofix",
]);
export const GLOBAL_DOCS_TREE: PageTree.Root = {
@@ -94,15 +96,17 @@ export const GLOBAL_DOCS_TREE: PageTree.Root = {
url: NESTED_DOCS_SECTIONS["build-agents"].entryUrl,
},
{ type: "separator", name: "Production" },
+ { type: "page", name: "Overview", url: "/docs/production" },
{
type: "page",
name: NESTED_DOCS_SECTIONS.gateway.title,
url: NESTED_DOCS_SECTIONS.gateway.entryUrl,
},
+ { type: "page", name: "Autofix", url: "/docs/autofix" },
{
type: "page",
- name: NESTED_DOCS_SECTIONS.observability.title,
- url: NESTED_DOCS_SECTIONS.observability.entryUrl,
+ name: NESTED_DOCS_SECTIONS.reliability.title,
+ url: NESTED_DOCS_SECTIONS.reliability.entryUrl,
},
{ type: "page", name: "Deploy your app", url: "/docs/deploy" },
{ type: "separator", name: "Reference" },
diff --git a/docs/next.config.mjs b/docs/next.config.mjs
index f901ef2dd..8d922913f 100644
--- a/docs/next.config.mjs
+++ b/docs/next.config.mjs
@@ -35,6 +35,21 @@ const config = {
},
async redirects() {
return [
+ {
+ source: "/docs/gateway/api/autofix",
+ destination: "/docs/autofix",
+ permanent: true,
+ },
+ {
+ source: "/docs/observability",
+ destination: "/docs/reliability",
+ permanent: false,
+ },
+ {
+ source: "/docs/observability/:path*",
+ destination: "/docs/reliability/:path*",
+ permanent: false,
+ },
// /cloud was the OpenUI Cloud product page; its pitch now lives on the
// home page, and the managed products have pages of their own. Temporary
// while the marketing restructure settles — make it permanent once the
diff --git a/docs/public/images/openui-architecture.svg b/docs/public/images/openui-architecture.svg
index d7e6db2cd..0359857a3 100644
--- a/docs/public/images/openui-architecture.svg
+++ b/docs/public/images/openui-architecture.svg
@@ -3,9 +3,9 @@
role="img"
aria-labelledby="openui-architecture-title openui-architecture-description"
>
- <title id="openui-architecture-title">How OpenUI turns a request into an interactive interface</title>
+ <title id="openui-architecture-title">Generation through Gateway</title>
<desc id="openui-architecture-description">
- A user request travels from an application to its backend, through OpenUI Gateway, and to a model provider. The corrected OpenUI Lang stream returns to a renderer, which displays the generated interface. Observability traces the flow.
+ A user request travels from an application to its backend, through OpenUI Gateway, and to a model provider. The corrected OpenUI Lang stream returns to a renderer, which displays the generated interface. Reliability Monitoring tracks generation failures.
</desc>
<defs>
@@ -142,8 +142,8 @@
<g>
<path d="M 890 590 H 1570" stroke="#fb7185" stroke-opacity="0.38" stroke-width="1.5" stroke-dasharray="5 7" fill="none" vector-effect="non-scaling-stroke" />
- <rect x="1010" y="555" width="220" height="70" rx="18" fill="#fb7185" fill-opacity="0.1" stroke="#fb7185" stroke-opacity="0.45" vector-effect="non-scaling-stroke" />
- <text x="1120" y="599" text-anchor="middle" fill="#fda4af" font-size="24" font-weight="500">Observability</text>
+ <rect x="970" y="555" width="300" height="70" rx="18" fill="#fb7185" fill-opacity="0.1" stroke="#fb7185" stroke-opacity="0.45" vector-effect="non-scaling-stroke" />
+ <text x="1120" y="599" text-anchor="middle" fill="#fda4af" font-size="24" font-weight="500">Reliability Monitoring</text>
</g>
</g>
</svg>
diff --git a/docs/public/images/openui-lang/reliability-layers.svg b/docs/public/images/openui-lang/reliability-layers.svg
index 910d57958..859368e31 100644
--- a/docs/public/images/openui-lang/reliability-layers.svg
+++ b/docs/public/images/openui-lang/reliability-layers.svg
@@ -9,7 +9,7 @@
<title id="title">Generative UI reliability pipeline</title>
<desc id="description">
A component schema and system prompt guide the selected model. The model produces generated UI,
- which is validated and corrected before reaching users. Observability measures the result and
+ which is validated and corrected before reaching users. Reliability Monitoring measures the result and
provides feedback for improvements across the pipeline.
</desc>
@@ -149,8 +149,8 @@
<rect class="observe" x="48" y="285" width="1024" height="78" rx="10" />
<rect class="accent-fill" x="48" y="285" width="6" height="78" rx="3" />
- <text class="observe-title ink" x="76" y="317">Observability</text>
- <text class="observe-copy muted" x="210" y="317">
+ <text class="observe-title ink" x="76" y="317">Reliability Monitoring</text>
+ <text class="observe-copy muted" x="270" y="317">
Measure errors, identify recurring failures, and verify improvements
</text>
<text class="observe-copy muted" x="76" y="344">