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

# Cloud vs Open Source

> Compare EverOS Cloud vs. Open Source to choose the right deployment for your applications.

## TL;DR (pick in 30 seconds)

<CardGroup cols={2}>
  <Card title="Choose EverOS Cloud" icon="cloud">
    Fastest path to production: use hosted APIs with managed reliability and minimal ops.
  </Card>

  <Card title="Choose Open Source" icon="github-alt">
    Full control, deploy on your own infrastructure, operate the stack—bring and maintain your own MongoDB, Milvus, Elasticsearch, and Redis.
  </Card>
</CardGroup>

**Not sure?** Start with EverOS Cloud for fast validation.

## Feature Comparison

<AccordionGroup>
  <Accordion title="Setup & getting started" description="Time to first stored memory and runtime requirements.">
    | Feature                      | EverOS Cloud           | Open Source                                        |
    | ---------------------------- | ---------------------- | -------------------------------------------------- |
    | **Time to get first memory** | 5 minutes              | 15 minutes                                         |
    | **Infrastructure predeploy** | None (managed service) | Required: MongoDB + Milvus + Elasticsearch + Redis |
    | **API Key setup**            | Managed on console     | No need                                            |
  </Accordion>

  <Accordion title="Core memory features" description="Key memory pipeline stages (from the EverOS framework).">
    | Feature                                              | EverOS Cloud | Open Source |
    | ---------------------------------------------------- | ------------ | ----------- |
    | **Semantic segmentation**                            | ✅            | ✅           |
    | **Episode generation (3rd-person narrative)**        | ✅            | ✅           |
    | **Atomic facts extraction**                          | ✅            | ✅           |
    | **Multi‑granularity indexing (Episode + Facts)**     | ✅            | ✅           |
    | **Thematic clustering**                              | ✅            | ✅           |
    | **Profile evolution (structured user profile)**      | ✅            | ✅           |
    | **Prospection generation (time‑scoped)**             | ✅            | ✅           |
    | **Hybrid retrieval (vector + BM25 + RRF)**           | ✅            | ✅           |
    | **Agentic retrieval (necessity & sufficiency loop)** | ✅            | ✅           |
    | **Evidence-based perception**                        | ✅            | ✅           |
  </Accordion>

  <Accordion title="Advanced capabilities" description="Production features beyond the core algorithm.">
    | Feature               | EverOS Cloud | Open Source                |
    | --------------------- | ------------ | -------------------------- |
    | **Auto‑scaling**      | ✅ Built-in   | Depends on your deployment |
    | **High availability** | ✅ Built-in   | Depends on your deployment |
    | **Dashboard**         | ✅ Built-in   | ❌                          |
    | **Memory visibility** | ✅ Built-in   | ❌                          |
  </Accordion>
</AccordionGroup>

## Next Steps

<CardGroup cols={3}>
  <Card title="Cloud overview" icon="cloud" href="/cloud/overview">
    Understand how EverOS Cloud ingests, organizes, and retrieves memories.
  </Card>

  <Card title="Cloud quickstart" icon="rocket" href="/cloud/quickstart">
    Make your first API call and save/retrieve memory in minutes.
  </Card>

  <Card title="Open source quickstart" icon="code" href="/open-source/quickstart">
    Set up the open-source stack and run EverOS in your own environment.
  </Card>
</CardGroup>

<Info>
  If you’re planning a production deployment and would like a recommendation, email us at <a href="mailto:contact@evermind.ai">[contact@evermind.ai](mailto:contact@evermind.ai)</a> with your constraints (region, compliance, scale, customization).
</Info>