> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.eyelevel.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.eyelevel.ai/_mcp/server.

# Choose Your GroundX Path

Pick the path that matches what you're trying to do. Each path leads to its own
guide; this page just points you there.

## Use Agent Harness

Use this when you want an agent (Claude, Codex, or another agent) to plan and
carry out GroundX work for you: load documents, check status, search, extract
structured data, debug a stuck process, or explain a GroundX concept, all with
GroundX-specific instructions instead of guesswork.

See [What Is GroundX Agent Harness?](/documentation/agent-harness/what-is-ground-x-agent-harness)
for an overview, then [Install GroundX Agent Harness](/documentation/agent-harness/install-ground-x-agent-harness) for
per-client installation steps.

Once it's installed, the agent reaches GroundX in one of two ways. It can call
GroundX's hosted tools, which you connect once and the agent then uses directly
(see [Connect Hosted MCP Tools](/documentation/agent-harness/connect-hosted-mcp-tools)). Or it can use the
GroundX SDK or REST API with your key, which works from any agent that can run
code or make web requests (see the [Direct SDK/API Quickstart](/documentation/get-started/direct-sdk-api-quickstart)). Connect the
hosted tools if your agent supports them; otherwise the agent uses the SDK.

## Use SDKs Or REST

Use this when you're building application code that calls GroundX directly: your
own service or backend, not an agent.

Start with the [Direct SDK/API Quickstart](/documentation/get-started/direct-sdk-api-quickstart), then see the
[GroundX SDKs](/sdks/ground-x-sd-ks) page for language-specific SDK
installation.

## Extract JSON From Documents

Use this when you want a document to come back as structured JSON your
application can use directly. You can do this through either Agent Harness or the
direct SDK/REST API; pick whichever fits how you want to work.

See [Extract Data from Documents](/documentation/guides/extract-data-from-documents)
for the direct SDK path, or [Use GroundX With Your Agent](/documentation/agent-harness/ingest-search-and-answer-with-an-agent#pull-structured-data-out-of-your-documents)
if you'd rather have an agent draft and iterate the extraction for you.

## Deploy GroundX On-Prem

Use this when you're planning a self-hosted or air-gapped GroundX deployment,
including an ingest-only mode (`mode: ingest`) that runs the ingest pipeline
without the retrieval/search stack.

See the **GroundX On-Prem** section, including [GroundX On-Prem on AWS](/documentation/ground-x-on-prem/ground-x-on-prem-on-aws) and [GroundX On-Prem on OpenShift](/documentation/ground-x-on-prem/ground-x-on-prem-on-open-shift).