Image-Text-to-Text
MLX
Safetensors
qwen3_5
clairic
cloudflare
systemone
structured-output
information-extraction
multimodal
custom-code
conversational
4-bit precision
Instructions to use mlx-community/clairic-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/clairic-4bit with MLX:
# Make sure mlx-vlm is installed # pip install --upgrade mlx-vlm from mlx_vlm import load, generate from mlx_vlm.prompt_utils import apply_chat_template from mlx_vlm.utils import load_config # Load the model model, processor = load("mlx-community/clairic-4bit") config = load_config("mlx-community/clairic-4bit") # Prepare input image = ["http://images.cocodataset.org/val2017/000000039769.jpg"] prompt = "Describe this image." # Apply chat template formatted_prompt = apply_chat_template( processor, config, prompt, num_images=1 ) # Generate output output = generate(model, processor, formatted_prompt, image) print(output) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Pi
How to use mlx-community/clairic-4bit with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "mlx-community/clairic-4bit"
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "mlx-lm": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "mlx-community/clairic-4bit" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent
How to use mlx-community/clairic-4bit with Hermes Agent:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "mlx-community/clairic-4bit"
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default mlx-community/clairic-4bit
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use mlx-community/clairic-4bit with OpenClaw:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "mlx-community/clairic-4bit"
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "mlx-community/clairic-4bit" \ --custom-provider-id mlx-lm \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
File size: 5,417 Bytes
1a48d6b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 | """Validation for embedded request images.
Callers translate ``EmbeddedImageError`` into their own validation error type.
The strict mode verifies the complete raster with Pillow; the signature mode
intentionally accepts header-only images for backends that validate only magic
bytes before forwarding the image to another service.
"""
from __future__ import annotations
import base64
import binascii
import io
import re
from collections.abc import Iterable
MAX_IMAGES = 4
MAX_IMAGE_BYTES = 4 * 1024 * 1024
MAX_IMAGES_BYTES = 8 * 1024 * 1024
_STRICT_DATA_URI = re.compile(
r"^data:(image/(?:png|jpe?g|webp));base64,([A-Za-z0-9+/]*={0,2})$", re.IGNORECASE
)
_DATA_URL = re.compile(r"^data:([^;,]+);base64,(.*)$", re.DOTALL)
_FORMATS = {"image/png": "PNG", "image/jpeg": "JPEG", "image/webp": "WEBP"}
class EmbeddedImageError(ValueError):
"""A request-image failure, with a stable code for the serving layer."""
def __init__(self, code: str, message: str) -> None:
super().__init__(message)
self.code = code
def normalize_embedded_image(
value: object,
*,
strict: bool = False,
max_image_bytes: int = MAX_IMAGE_BYTES,
max_pixels: int = 16_000_000,
) -> dict[str, str]:
"""Normalize a data URL or exact ``{content_type, base64}`` object.
``strict=True`` verifies restricted data-URI syntax and complete raster data
with Pillow, including the dimension limit. The default checks image
signatures and accepts a more permissive data-URL syntax.
"""
if isinstance(value, str):
matched = (_STRICT_DATA_URI if strict else _DATA_URL).fullmatch(value)
if matched is None:
label = "URI" if strict else "URL"
raise EmbeddedImageError(
"invalid_image", f"image must be an embedded base64 data {label}"
)
value = {"content_type": matched.group(1), "base64": matched.group(2)}
if type(value) is not dict or set(value) != {"content_type", "base64"}:
raise EmbeddedImageError("invalid_image", "image object has invalid fields")
content_type = value["content_type"]
encoded = value["base64"]
if type(content_type) is not str or type(encoded) is not str:
raise EmbeddedImageError("invalid_image", "image fields must be strings")
canonical_type = content_type.lower().replace("image/jpg", "image/jpeg")
if canonical_type not in _FORMATS:
raise EmbeddedImageError("invalid_image", "image content type is not supported")
try:
decoded = base64.b64decode(encoded, validate=True)
except (binascii.Error, ValueError) as error:
raise EmbeddedImageError(
"invalid_image", "image base64 is malformed"
) from error
if not decoded:
raise EmbeddedImageError("invalid_image", "image payload is empty")
if len(decoded) > max_image_bytes:
raise EmbeddedImageError(
"image_too_large", "image exceeds the decoded byte limit"
)
if strict:
_verify_raster(decoded, _FORMATS[canonical_type], max_pixels=max_pixels)
elif not _has_image_signature(canonical_type, decoded):
raise EmbeddedImageError(
"invalid_image", "image payload does not match its content type"
)
return {
"content_type": canonical_type,
"base64": base64.b64encode(decoded).decode("ascii"),
}
def _verify_raster(data: bytes, expected_format: str, *, max_pixels: int) -> None:
from PIL import Image, UnidentifiedImageError
try:
with Image.open(io.BytesIO(data)) as image:
if image.format != expected_format:
raise EmbeddedImageError(
"invalid_image", "image payload does not match its content type"
)
if image.width * image.height > max_pixels:
raise EmbeddedImageError(
"invalid_image", "image dimensions are too large"
)
image.verify()
except EmbeddedImageError:
raise
except (
UnidentifiedImageError,
OSError,
ValueError,
Image.DecompressionBombError,
) as error:
raise EmbeddedImageError("invalid_image", "image payload is invalid") from error
def _has_image_signature(content_type: str, data: bytes) -> bool:
if content_type == "image/png":
return data.startswith(b"\x89PNG\r\n\x1a\n")
if content_type == "image/jpeg":
return data.startswith(b"\xff\xd8\xff")
return len(data) >= 12 and data.startswith(b"RIFF") and data[8:12] == b"WEBP"
def decoded_base64_size(encoded: str) -> int:
"""Size of already validated, canonical base64 without decoding again."""
return len(encoded) * 3 // 4 - (len(encoded) - len(encoded.rstrip("=")))
def validate_image_count(value: object, *, max_images: int = MAX_IMAGES) -> object:
if not isinstance(value, (list, tuple)):
raise EmbeddedImageError("invalid_image", "images must be an array")
if len(value) > max_images:
raise EmbeddedImageError(
"too_many_images", "request exceeds the image count limit"
)
return value
def validate_total_image_bytes(
sizes: Iterable[int], *, max_bytes: int = MAX_IMAGES_BYTES
) -> None:
if sum(sizes) > max_bytes:
raise EmbeddedImageError(
"images_too_large", "images exceed the aggregate decoded byte limit"
)
|