Cloudflare

Cloudflare Workers AI provides edge-based embedding inference with BGE and other open-source models.

Overview

Models: 7 models (BGE variants, Qwen3, others)
Key Features: Edge inference, low latency, open-source models
API Docs: Cloudflare Workers AI

Environment Variables

export CLOUDFLARE_API_KEY="your-cloudflare-api-key"
export CLOUDFLARE_ACCOUNT_ID="your-account-id"

Supported Parameters

Parameter	Type	Required	Description
`model`	str	Yes	Model identifier (e.g., `@cf/baai/bge-base-en-v1.5`)
`input`	str \| List[str]	Yes	Text(s) to embed
`pooling`	str	No	`"mean"` or `"cls"` (for BGE models)
`api_key`	str	No	Override API key
`account_id`	str	No	Override account ID

Note: Cloudflare does not support input_type or dimensions parameters.

Examples

Basic Usage

response = client.embed(
    model="@cf/baai/bge-base-en-v1.5",
    input="Hello, Cloudflare!"
)

With Account ID

client = catsu.Client(
    api_keys={
        "cloudflare": "your-api-key"
    }
)

response = client.embed(
    model="@cf/baai/bge-base-en-v1.5",
    input="Text",
    account_id="your-account-id"  # If not in env
)

With Pooling (BGE models)

# Mean pooling (default)
response = client.embed(
    model="@cf/baai/bge-base-en-v1.5",
    input="Text",
    pooling="mean"
)

# CLS token pooling
response = client.embed(
    model="@cf/baai/bge-base-en-v1.5",
    input="Text",
    pooling="cls"
)

Special Notes

⚠️ input_type and dimensions are NOT supported
Requires CLOUDFLARE_ACCOUNT_ID in addition to API key
Edge-based inference for low latency
BGE models support pooling parameter
Model names use @cf/ prefix

Next Steps

Models Catalog - View all Cloudflare models

On this page