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

# When to Use Monid Directly

> Use Monid directly when it's an internal building block of your product, not something users see.

Use Monid directly when **your users never see Monid** — it's an internal building block powering features inside your product. You hold a single Monid account and call the API server-to-server.

## Use Direct Integration When

* Monid is a hidden implementation detail (data enrichment, background jobs, agentic tools).
* You want the simplest possible integration — one API key, no OAuth flow.
* You bill users through your own system (subscription, per-feature, or free).
* You want a **unified billing surface**: one Monid invoice for all usage across your users.

## Two Patterns

<CardGroup cols={2}>
  <Card title="Full discover / inspect / run" icon="compass">
    Let your backend (or an internal agent) pick endpoints dynamically at runtime. Best when the set of tools needed isn't known ahead of time.
  </Card>

  <Card title="Curated endpoints only" icon="list-check">
    Pre-select a handful of endpoints and call `run` on them directly. Best when you want a simple, fixed interface with predictable pricing.
  </Card>
</CardGroup>

## Skip Direct Integration When

* Your users should have their own Monid account and wallet → use [OAuth](/integrations/oauth/when-to-use).
* You want to expose Monid tools to user-owned agents but with your own billing and allowlist → use [a Proxy](/integrations/proxy/when-to-use).

## Examples

<AccordionGroup>
  <Accordion title="CRM that auto-enriches new leads">
    Whenever a new contact is created, the CRM's backend calls a curated Monid endpoint to pull LinkedIn + company data. Users never see "Monid" — they just see enriched contacts. The CRM pays Monid once a month and bills users on its own plan.
  </Accordion>

  <Accordion title="Agentic feature inside a SaaS product">
    A support tool has a "Research this ticket" button. Its internal agent uses Monid `discover` / `inspect` / `run` to find the right data source for each ticket. The user only sees the final summary.
  </Accordion>

  <Accordion title="Batch pipeline for analytics">
    A nightly job runs a fixed set of Monid scraping endpoints to refresh a dataset. No user interaction, no dynamic discovery — just direct `run` calls on known endpoints.
  </Accordion>
</AccordionGroup>

Ready to build? Continue to [Set Up Direct Access](/integrations/direct).