Run Marker app, API server, client, and deployment surfaces safely.
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill server-deployment --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Server Deployment?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-server-deployment)More formats (shields.io, HTML) on the badges page.
---
name: server-deployment
description: "Run Marker app, API server, client, and deployment surfaces safely."
disable-model-invocation: true
metadata:
disco-role: operating
license: NOASSERTION
---
# Marker Server Deployment
Use this sub-skill when the task is to run, call, or adapt Marker’s local app/server/deployment surfaces rather than choose the core conversion pipeline.
## Choose the surface
- **Local FastAPI server**: use `marker_server --host HOST --port PORT`; read [apps and server](references/apps-and-server.md) and [API contract](references/api-contract.md).
- **Client integration**: use [scripts/marker_server_client_template.py](scripts/marker_server_client_template.py) for explicit `/marker` filepath or `/marker/upload` multipart requests.
- **Interactive apps**: use `marker_gui` for document-to-markdown/JSON/HTML/chunks trials and `marker_extract` for structured extraction UI; read [apps and server](references/apps-and-server.md).
- **Modal deployment**: adapt the reference-only Modal pattern in [modal deployment](references/modal-deployment.md) when a hosted GPU endpoint is needed.
- **Smoke checks**: run [scripts/server_cli_smoke.py](scripts/server_cli_smoke.py) to verify server CLI help/imports without starting a long-running service.
## Boundaries and routes
- For `output_format` semantics, page ranges, table/OCR converter choices, image extraction, and CLI/Python conversion workflows, route to sibling sub-skill `../conversion-cli-api/`.
- For `use_llm`, service selection, credentials, and structured extraction converter behavior outside the Streamlit launcher, route to sibling sub-skill `../llm-extraction-services/`.
- For custom processors, renderers, converter classes, config JSON, and debug internals, route to sibling sub-skill `../configuration-extension/`.
- Keep deployment checks safe: do not start `marker_server`, `marker_gui`, `marker_extract`, Modal deploys, model downloads, or external network calls unless the user explicitly asks.
## Minimum safe commands
```bash
python scripts/server_cli_smoke.py --check-import
python scripts/marker_server_client_template.py --help
```
Install only the optional surface you need: FastAPI dependencies for `marker_server`, Streamlit dependencies for app launchers, Modal dependencies for cloud deployment, and `marker-pdf[full]` only when non-PDF documents are in scope.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!