dcc-mcp-core-semantic
Native Rust semantic embeddings for dcc-mcp-core,
shipped as a separate PyPI wheel so the main dcc-mcp-core install stays
free of ONNX Runtime and the ~25-40 MB wheel size that comes with it.
When to install this
Only when you actually need dense semantic recall for skill / capability
search. The default pip install dcc-mcp-core install ships with a
HashedEmbedder (zero-dep, hashing-trick + character n-grams) which is
enough for ≤100-skill DCC adapters and tolerates morphology variants
like render / rendering already.
If your skill catalogue grows to many hundreds of skills, or your agents
ask in natural language that does not share token structure with your
SKILL.md metadata, install this companion wheel to upgrade
OnnxEmbedder to true dense semantic recall.
How to install
pip install 'dcc-mcp-core[semantic]'
This pulls in dcc-mcp-core-semantic (this package) via the [semantic]
extra, plus fastembed as a Python-side fallback for platforms where the
Rust wheel is not yet available.
You can also install this package directly if you want only the Rust backend without the Python fallback:
pip install dcc-mcp-core dcc-mcp-core-semantic
How it gets used
Your adapter code does not change. Once installed,
dcc_mcp_core.OnnxEmbedder() automatically prefers the Rust extension:
from dcc_mcp_core import OnnxEmbedder, VectorSkillIndex
# Loads the BAAI/bge-small-en-v1.5 model on first use, cached to
# ~/.cache/fastembed/ (or wherever DCC_MCP_EMBED_MODEL_DIR points).
emb = OnnxEmbedder()
idx = VectorSkillIndex(embedder=emb)
Configuration
Both env vars are honoured by OnnxEmbedder regardless of which backend
serves the call:
| Variable | Default | Purpose |
|---|---|---|
DCC_MCP_EMBED_MODEL |
BAAI/bge-small-en-v1.5 |
HuggingFace model name. Must be one of dcc_mcp_core_semantic.native.SUPPORTED_MODELS. |
DCC_MCP_EMBED_MODEL_DIR |
unset (fastembed default) | On-disk cache for the ONNX model bytes. Pre-place this on a shared mount for firewalled studios. |
Build from source
The source lives in the main dcc-mcp-core repository, NOT here.
Rust crate: crates/dcc-mcp-semantic/. Wheel build:
cd pkg/dcc-mcp-core-semantic
maturin build --release
ONNX Runtime is pulled in at build time via the ort crate's
download-binaries strategy. The wheel that maturin produces is
self-contained — end users do not need to install ONNX Runtime separately.
License
MIT, matching dcc-mcp-core.
1 maintainer
Release files for dcc-mcp-core-semantic 0.20.30
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Built distributions (wheels)
| File | Reset | |||
|---|---|---|---|---|
| dcc_mcp_core_semantic-0.20.30-cp38-abi3-win_amd64.whl | CPython 3.8 | abi3 | Windows x86-64 | Details |
| dcc_mcp_core_semantic-0.20.30-cp38-abi3-manylinux_2_28_x86_64.whl | CPython 3.8 | abi3 | Linux glibc 2.28+ x86-64 | Details |
| dcc_mcp_core_semantic-0.20.30-cp38-abi3-macosx_11_0_arm64.whl | CPython 3.8 | abi3 | macOS 11.0+ ARM64 | Details |
| dcc_mcp_core_semantic-0.20.30-cp37-cp37m-win_amd64.whl | CPython 3.7 | CPython 3.7 pymalloc | Windows x86-64 | Details |
| dcc_mcp_core_semantic-0.20.30-cp37-cp37m-manylinux_2_28_x86_64.whl | CPython 3.7 | CPython 3.7 pymalloc | Linux glibc 2.28+ x86-64 | Details |
Total release size: 51.7 MB
Release files / dcc_mcp_core_semantic-0.20.30-cp38-abi3-win_amd64.whl
| Download URL | dcc_mcp_core_semantic-0.20.30-cp38-abi3-win_amd64.whl |
|---|---|
| Size | 10.3 MB |
| Tags | CPython 3.8 Windows x86-64 abi3 |
|
SHA-256 checksum How to use checksums |
5441d00abe8bd85d62feaa903ebc1d6ef740e5eb16630a6c7b323eac320a80e2
|
|
BLAKE2b-256 checksum How to use checksums |
64cde9493c452976f9e019ce5662cd46fba3398d66d8c38b1bb16789f478de4c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency logRelease files / dcc_mcp_core_semantic-0.20.30-cp38-abi3-manylinux_2_28_x86_64.whl
| Download URL | dcc_mcp_core_semantic-0.20.30-cp38-abi3-manylinux_2_28_x86_64.whl |
|---|---|
| Size | 10.8 MB |
| Tags | CPython 3.8 Linux glibc 2.28+ x86-64 abi3 |
|
SHA-256 checksum How to use checksums |
d26c92377cf8b2e573474d117be3cc7a04fff75949f8c623e62555e1c93985f8
|
|
BLAKE2b-256 checksum How to use checksums |
2bca54ab2573e1a09daed5de6876c3e286717a4e5e1ee3ea1e8f76cc8543b4fb
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency logRelease files / dcc_mcp_core_semantic-0.20.30-cp38-abi3-macosx_11_0_arm64.whl
| Download URL | dcc_mcp_core_semantic-0.20.30-cp38-abi3-macosx_11_0_arm64.whl |
|---|---|
| Size | 9.6 MB |
| Tags | CPython 3.8 abi3 macOS 11.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
24a3c912102847596e609e6188058ad2a7fa063231e227b18533b7fe49b0898f
|
|
BLAKE2b-256 checksum How to use checksums |
3efbd33f7b9118f72d7f0e3088563e1318e570356574d02a15e86d11a7558018
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency logRelease files / dcc_mcp_core_semantic-0.20.30-cp37-cp37m-win_amd64.whl
| Download URL | dcc_mcp_core_semantic-0.20.30-cp37-cp37m-win_amd64.whl |
|---|---|
| Size | 10.3 MB |
| Tags | CPython 3.7 CPython 3.7 pymalloc Windows x86-64 |
|
SHA-256 checksum How to use checksums |
e6a5bf4f87c3c960e28ef6d84f8d23107c18264a5db31cb16eb9ff04c7b4c06a
|
|
BLAKE2b-256 checksum How to use checksums |
f67a1752be4ba8504ed1ccfc996ff72ff5cb1e473f09ba2e3b9701c1fb060b8c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency logRelease files / dcc_mcp_core_semantic-0.20.30-cp37-cp37m-manylinux_2_28_x86_64.whl
| Download URL | dcc_mcp_core_semantic-0.20.30-cp37-cp37m-manylinux_2_28_x86_64.whl |
|---|---|
| Size | 10.8 MB |
| Tags | CPython 3.7 CPython 3.7 pymalloc Linux glibc 2.28+ x86-64 |
|
SHA-256 checksum How to use checksums |
930fdad19cdd03b7326c60dcaef2cb90af9cf6a1aac6446dd1a5f4a5ab322be2
|
|
BLAKE2b-256 checksum How to use checksums |
8bcaa6aaf1c2483e073e28ca17d07c53506f2390d562ffdbabdbbeaff331fe64
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 18, 2026.
Transparency log
