AlbumentationsX MCP

Model Context Protocol server for AlbumentationsX: inspect datasets, preview augmentations, refine them with visual feedback, and export reproducible pipelines.

Baseline and adjusted AlbumentationsX preview contact sheets

Ask an MCP host for several robustness variants, reject an excessive result such as too_noisy:high, compare the adjusted batch previews, and export the accepted pipeline.

Install

Claude Desktop

Download the latest albumentationsx-mcp.mcpb, install it from Settings -> Extensions -> Advanced settings, and select separate image and artifact directories.

Other MCP Hosts

Run the published server with bounded local access:

uvx --from albumentationsx-mcp albumentationsx-mcp \
  --allowed-root /absolute/path/to/images \
  --artifact-root /absolute/path/to/albu-artifacts

full is the v1.x default. Add --capability-profile review for a preview-focused tool surface; see configuration. Copyable host configurations are in the install guide. The repository also contains a native Codex plugin bundle. npx skills add dKosarevsky/albu-mcp installs agent guidance, not the MCP server.

First Preview

After connecting the server, ask your host:

Use AlbumentationsX MCP on /absolute/path/to/images.
Run the smoke check, start with a low-intensity pipeline, validate the request,
render one variant per image, and show me the contact sheet before exporting anything.

Read albumentationsx://examples/client-smoke; if resource reads are unavailable, call get_workflow_example with example_id="client-smoke".
Call run_host_smoke_check; continue only when preview_ready is true, using its preview_request_template.
Call validate_preview_request before rendering and compare preview runs before accepting a candidate.
Give concrete feedback such as too_noisy:high or exposure_too_weak:medium, then export the final pipeline.

If setup fails, read albumentationsx://diagnostics/guide and call diagnose_environment for bounded remediation actions.

Capabilities

Transform discovery, schemas, recipes, and pipeline validation.
Classification, detection, segmentation, OCR, bbox, mask, keypoint, and dataset-quality workflows.
Deterministic previews, contact sheets, annotation overlays, comparison, ranking, and reports.
Interactive MCP Apps review with a text-only fallback for other hosts.
Structured feedback, tuning sessions, and Python, JSON, or YAML export.
Stable agent workflow resources, prompts, diagnostics, and reviewed contract snapshots.

The server does not execute arbitrary Python, fetch remote images, overwrite datasets, or train models. Reads are restricted by --allowed-root; generated files stay under --artifact-root.

Integrations

Documentation

Install and host configuration
Runtime settings and capability profiles
First 10 minutes
Usage and recipes
MCP Apps review and compatibility policy
Documentation index
CHANGELOG.md
server.json: public MCP Registry metadata.

Development

uv sync --all-extras --dev
uv run pytest
uv run ruff check .
uv run ruff format --check .
uv run ty check

Licensed under AGPL-3.0-or-later.

Albu MCP

README