> ## Documentation Index
> Fetch the complete documentation index at: https://docs.endprompt.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Welcome to Endprompt

> Build production-ready LLM-powered APIs in minutes

## What is Endprompt?

Endprompt is a platform for creating, managing, and deploying LLM-powered API endpoints. Instead of wrestling with prompt engineering, versioning, and infrastructure, you focus on building great products.

<CardGroup cols={2}>
  <Card title="Create Endpoints" icon="plug">
    Define stable API paths with typed input and output schemas
  </Card>

  <Card title="Write Prompts" icon="wand-magic-sparkles">
    Use Liquid templating to create dynamic, reusable prompts
  </Card>

  <Card title="Test & Iterate" icon="flask">
    Test prompts instantly with auto-generated forms or bulk CSV uploads
  </Card>

  <Card title="Deploy with Confidence" icon="rocket">
    Version prompts, promote to live, and rollback safely
  </Card>
</CardGroup>

## Why Endprompt?

Most teams building with LLMs face the same challenges:

* **No versioning discipline** — Prompts change without history or rollback capability
* **Inconsistent schemas** — Input/output formats drift over time
* **No observability** — Hard to debug why a response was wrong
* **Painful iteration** — Testing changes requires code deployments

Endprompt solves these with:

<AccordionGroup>
  <Accordion title="Stable API Endpoints" icon="link">
    Your integration code never changes. Swap prompts, models, and configurations behind a stable URL like `https://yourcompany.api.endprompt.ai/api/v1/summarize`.
  </Accordion>

  <Accordion title="Strong Schema Enforcement" icon="shield-check">
    Define exactly what inputs your endpoint accepts and what outputs it returns. Validation happens automatically—malformed requests are rejected before hitting the LLM.
  </Accordion>

  <Accordion title="Safe Iteration Workflow" icon="git-alt">
    Draft prompts, test them, promote to live. Keep multiple versions and rollback instantly if something goes wrong.
  </Accordion>

  <Accordion title="Built-in Observability" icon="chart-line">
    Every execution is logged with inputs, outputs, latency, token usage, and cost. Replay any request to debug issues.
  </Accordion>
</AccordionGroup>

## How It Works

<Steps>
  <Step title="Create an Endpoint">
    Define your API path (e.g., `/api/v1/summarize`) and specify the input fields your endpoint accepts.
  </Step>

  <Step title="Write a Prompt">
    Create a prompt using Liquid templating. Reference your input fields with `{{ inputs.fieldName }}`.
  </Step>

  <Step title="Test Your Prompt">
    Use the built-in test runner to send requests and see real LLM responses instantly.
  </Step>

  <Step title="Go Live">
    Promote your prompt to "Live" status. It's now the default for your endpoint.
  </Step>

  <Step title="Call Your API">
    Make HTTP requests to your endpoint URL with your API key. That's it!
  </Step>
</Steps>

## Ready to Start?

<CardGroup cols={2}>
  <Card title="Quickstart Guide" icon="bolt" href="/quickstart">
    Build your first endpoint in 5 minutes
  </Card>

  <Card title="How Endprompt Works" icon="circle-info" href="/getting-started/how-it-works">
    Understand the core concepts
  </Card>
</CardGroup>