orca-profiles-mcp

orca-profiles-mcp

An MCP server that reads and edits OrcaSlicer profiles, expanding the full inheritance chain and tracking the provenance of every value.

Category
Visit Server

README

orca-profiles-mcp

An MCP server that reads and edits OrcaSlicer profiles. It expands the whole inherits chain and shows which link set every value.

What it does

An OrcaSlicer profile stores only its differences from its parent. Opening the file is not enough to learn the outer wall speed: the value may come from the parent, from the parent's parent, or from the engine's built-in defaults. This server walks the chain to its root, computes the effective values and keeps their provenance.

It handles the parts of Orca that break naive readers:

  • parents are resolved by name, across vendors — a Sovol filament inherits from the OrcaFilamentLibrary;
  • a missing Generic … parent falls back to Generic <material> @System, and renamed profiles are found through their renamed_from list;
  • vectors bound to extruder variants merge element by element, including the nil marker that means "keep the parent's value here";
  • a child may declare more extruders than its parent — a toolchanger on top of a single-extruder machine profile — and both extruders keep their values;
  • keys that do not belong to the profile's type are dropped, because Orca drops them too; they are reported instead of being silently applied.

Setup

cd ~/job/orca-profiles-mcp
uv sync

Register with Claude Code:

claude mcp add orca-profiles -- uv --directory ~/job/orca-profiles-mcp run orca-profiles-mcp

Tools

Tool Purpose
get_setup discovered roots, Orca version, vendors, selected presets
list_profiles search by type, name, vendor, source
get_profile profile in raw / resolved / traced mode
get_chain inheritance chain to the root
explain_key where a specific value came from
find_children which profiles inherit from this one
diff_profiles compare two profiles
compare_with_upstream compare against the OrcaSlicer repository
validate broken inherits, cycles, unknown keys, redundant deltas
check_deltas verify expansion against the deltas Orca itself wrote
reports redundant keys and unexplained vector lengths separately
set_values set values with delta recomputation
create_profile create a user profile
rename_profile, delete_profile rename and delete
normalize_profile drop keys identical to the parent's

Values are returned exactly as the files store them: "0.20" stays "0.20", and "100%" stays a percentage. The engine normalises both when it slices; this server does not, so what you read is what the profile says and what gets written back is unchanged.

Writing profiles

Your own profiles are edited freely. Touching a system or bundled one requires force=true: those files belong to a vendor library, the next profile update restores them, and every descendant inherits the change.

Writes go through a temporary file and an atomic rename, so an interrupted write cannot truncate a profile, and a backup copy is made next to the file first (backup=false disables it). Values are checked before anything is written — a key belonging to another profile type, a vector where the engine wants a scalar, or a number where the format requires a string is refused rather than stored and silently dropped by Orca. Settings your pinned snapshot does not recognise are carried through untouched, so a profile written by a newer Orca does not lose them.

OrcaSlicer reads profiles at startup and rewrites them at exit. Edits made while the application is running will be overwritten — close Orca first.

How correctness is checked

Orca's CLI cannot verify inheritance. --export-settings returns the engine defaults plus the keys present in the files handed to it and never walks inherits; selecting presets through OrcaSlicer.conf does not activate them; passing a whole chain as a file list is rejected as a duplicate config file. All three were tried against 2.4.2.

What Orca does leave behind is evidence of its own expansion: when it saves a user profile it stores the minimal delta against the fully expanded parent. So every profile Orca has written is a recorded answer. check_deltas recomputes those deltas and compares — a mismatch means the resolver expands a parent differently than the engine did.

This is what tests/test_smoke.py asserts against the installed library, and it is how the toolchanger vector defect was found.

Keys stored despite already matching the parent are reported separately: they are harmless leftovers from before a parent changed, not resolver errors.

Refreshing the engine snapshot

Engine defaults, the variant key sets and the per-type key lists are captured from the installed Orca into src/orca_profiles_mcp/data/engine-snapshot.json, pinned to its version tag. After upgrading OrcaSlicer:

uv run python scripts/build_snapshot.py
uv run pytest

Tests

uv run pytest

Four layers:

  • unit tests over a miniature fixture library, one per edge case — cross-vendor inheritance, the Generic fallback, renamed_from, cycles, missing parents, stride-1 and stride-2 vectors, nil elements;
  • write round-trips: read → write unchanged → the file is byte-identical;
  • end-to-end tests that spawn the packaged entry point as a subprocess and drive it over MCP stdio, covering tool registration, a full create/edit/verify/delete cycle, and error propagation;
  • checks against the installed library, including the delta verification described above.

Tests using the real library skip themselves when no Orca data directory is present; everything else runs against fixtures.

Documents

  • Design: docs/specs/2026-08-17-orca-profiles-mcp-design.md — includes what reading the OrcaSlicer sources established about the profile format, and what the CLI experiments ruled out
  • Implementation plan: docs/plans/2026-08-17-orca-profiles-mcp.md

Licence

AGPL-3.0, matching OrcaSlicer itself. This project contains no OrcaSlicer code, but its inheritance logic is a line-by-line port of Preset.cpp and PrintConfig.cpp, and the engine snapshot holds default values and key lists extracted from that source.

Not affiliated with or endorsed by the OrcaSlicer project.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured