chp-adapter-mlx
Apple Silicon native text generation as governed CHP capabilities, backed by a
local mlx_lm server. MLX is the fastest
local inference path on Apple Silicon (≈2–3× Ollama, ≈1.5× llama.cpp).
mlx_lm.server is OpenAI-compatible, so this adapter mirrors chp-adapter-vllm /
chp-adapter-local-llm and the gateway routes it as another inference owner for
capacity-aware (GPU-utilization) routing.
Capabilities
| Capability | Risk | Description |
|---|---|---|
chp.adapters.mlx.status |
low | Is mlx/mlx-lm installed (+ versions) and is the server reachable? |
chp.adapters.mlx.list_models |
low | Models served by mlx_lm.server (/v1/models) |
chp.adapters.mlx.generate |
medium | Single-turn completion (/v1/completions) |
chp.adapters.mlx.chat |
medium | Multi-turn chat (/v1/chat/completions) |
Composition & evidence
The adapter imports no HTTP library — every server call routes through
chp.adapters.http, so HTTP is its own governed evidence chain and the adapter is
conformance-clean. Prompt/completion/message content is never recorded in
evidence; only model id, token counts, latency, and errors are.
Running the backend (inference node)
pip install mlx-lm
# OpenAI-compatible server on :8081 (8080 collides with llama.cpp)
mlx_lm.server --model mlx-community/Qwen3-... --port 8081
Config via env: MLX_BASE_URL (default http://localhost:8081), MLX_MODEL,
MLX_API_KEY. Add "mlx" to the host profile's adapters list to register it.
Verify over the mesh with chp.adapters.mlx.status.
1 maintainer
Metadata
1 maintainer
Release files for chp-adapter-mlx 0.54.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| chp_adapter_mlx-0.54.0.tar.gz | 21.5 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| chp_adapter_mlx-0.54.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 40.9 kB
Release files / chp_adapter_mlx-0.54.0.tar.gz
| Download URL | chp_adapter_mlx-0.54.0.tar.gz |
|---|---|
| Size | 21.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
4802c80f949fd537ee3bf26539a37c5509a0d13d4ed658a4e59daeeef8523c9b
|
|
BLAKE2b-256 checksum How to use checksums |
e7e9cc2e3b94f95e1224416f5cfdd22e9e90e15e71b81f2ab244e98c92c3ec49
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.7
|
Release files / chp_adapter_mlx-0.54.0-py3-none-any.whl
| Download URL | chp_adapter_mlx-0.54.0-py3-none-any.whl |
|---|---|
| Size | 19.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
a88f0cb5c2d68a94c04109b0c9759263ed177686e7a23bcbbde1b24c020c66a8
|
|
BLAKE2b-256 checksum How to use checksums |
68db7359905ba93250aa7a57c6d16addfb52e48450cb2c3f59791ab5b3713236
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.7
|
Security for chp-adapter-mlx 0.54.0
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.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
