Native Rust command-line browser automation tool built for AI agents, with compact accessibility snapshots and 50+ commands.
Open source
Cloudflare Workers AI
Cloudflare Workers AI runs 50+ open-source AI models on serverless GPUs, billed per use in Neurons from Workers or the REST API.
What is Cloudflare Workers AI?
Workers AI is Cloudflare's serverless inference service. You call models running on GPUs in Cloudflare's network from a Worker, from Pages, or from anywhere through the Cloudflare API, without renting or scaling GPUs yourself. The catalog covers more than 50 open-source models for text generation, embeddings, image classification, object detection, speech recognition, translation, summarization and text-to-image, and features such as function calling, JSON Mode, prompt caching, fine-tunes and an asynchronous Batch API are documented. Billing is based on Neurons, a unit that measures the GPU compute behind a request, and each model page maps tokens, audio seconds or images to Neurons. Workers AI is included in the Free and Paid Workers plans with a free allocation of 10,000 Neurons per day, and usage beyond that on Workers Paid costs $0.011 per 1,000 Neurons. Some larger models need a paid billing method, and prepaid AI Gateway credits can also pay for inference.
Use cases
- Run open-source LLMs from a Worker
- Generate text embeddings for search
- Transcribe audio with speech recognition
- Generate images from text prompts
- Classify images and detect objects
- Translate and summarize text via API
Who it's for
Why use Cloudflare Workers AI
- There is no GPU fleet to rent, patch or scale, and you pay only for the inference you use.
- The same platform covers models, vector search, gateways and compute, so an AI feature can live in one Worker.
- Free allocation lets developers prototype before adding a paid plan.
- Per-model pricing shows what each model costs in tokens or other units and in Neurons.
- Custom requirements such as private custom models or higher limits can be requested through a form.
Cloudflare Workers AI features
- 50+ open-source models in a catalog
- Serverless GPU inference with no servers to manage
- Access from Workers, Pages or the REST API
- Function calling and JSON Mode
- Prompt caching and fine-tunes
- Asynchronous Batch API
- Markdown conversion
- Per-model Neuron-based pricing with a usage dashboard
- Prepaid AI Gateway credits with Unified billing
Cloudflare Workers AI integrations
How to get started with Cloudflare Workers AI
- Create a Cloudflare account and open the Workers AI section of the dashboard.
- Pick a model from the catalog and add an AI binding to a Worker, or get an API token for the REST API.
- Send a prompt, image, audio file or text to the model and read the response.
- Monitor Neuron usage in the dashboard.
- Move to Workers Paid, or load AI Gateway credits, when you pass the daily free allocation.
Cloudflare Workers AI limitations
- Free allocation is 10,000 Neurons per day, reset at 00:00 UTC
- Operations fail with an error once a daily limit is exceeded
- Text generation defaults to 300 requests per minute
- Models flagged as paid are limited to 20 requests per minute on standard billing
- Some models need Workers Paid or prepaid AI Gateway credits
Cloudflare Workers AI pricing
Yes, Cloudflare Workers AI has a free plan. Paid plans start at $0.01 usage-based, according to the official pricing page on .
Cloudflare Workers AI plans
Workers Free
Free
10,000 Neurons per day
- 10,000 Neurons per day at no charge
- Upgrade to Workers Paid for more usage
Workers Paid
$0.011 /1,000 Neurons
above the 10,000 daily free Neurons
- 10,000 Neurons per day free allocation
- Pay only for usage beyond the allocation
- Access to models that require paid billing
Plans as listed on the official pricing page on . Prices can change and may vary by country or exclude tax. Check the official pricing page.
Cloudflare Workers AI alternatives
The main Cloudflare Workers AI alternatives listed on Debuly are Replicate, Together AI, Fireworks AI, Modal, DeepInfra and FlexAI. The table compares pricing and platforms; order is A–Z.
| Tool | Pricing | Starting price | Platforms |
|---|---|---|---|
| Cloudflare Workers AI | Freemium | $0.01, usage-based | API, Web |
| DeepInfra | Paid | — | API, Web |
| Fireworks AI | Paid | $0.01, usage-based | API, Web |
| FlexAI | Paid | $0.01, usage-based | API |
| Modal | Freemium | $250/mo | Web, API, CLI, Python |
| Replicate | Paid | $0.09, usage-based | API, Web, Python, Node.js |
| Together AI | Paid | — | Web, API |
DeepInfra
AI inference cloud with an OpenAI-compatible API for 100+ open models, plus private GPU deployments, GPU rental and hosted agents.
Fireworks AI
Inference and training platform for open models: pay-per-token serverless APIs, GPU deployments, fine-tuning and reserved capacity.
FlexAI
OpenAI-compatible inference API, dedicated GPU endpoints and fine-tuning for open models, billed per token or per GPU-hour.
Modal
Serverless cloud platform for running AI inference, training, batch jobs and agent sandboxes, billed per second of compute.
FreemiumVisit Modal (official site)
Replicate
Cloud API for running, fine-tuning and deploying open-source and proprietary AI models, billed by compute time or by output.
Together AI
AI cloud for running open-source models through serverless and batch inference, dedicated endpoints, GPU clusters and fine-tuning.
Cloudflare Workers AI FAQ
What is Workers AI?
Workers AI is Cloudflare's serverless service for running AI models on GPUs in its global network. You call it from a Worker, from Pages or through the REST API.
How much does Workers AI cost?
Usage above the free allocation is billed at $0.011 per 1,000 Neurons on the Workers Paid plan. Each model's page translates this into per-token or per-unit prices.
Is there a free plan for Workers AI?
Yes. Both Workers Free and Workers Paid include 10,000 Neurons per day at no charge. Workers Free cannot go beyond that and must upgrade to Workers Paid.
What are Neurons in Workers AI?
Neurons are Cloudflare's unit for measuring AI output across models. They represent the GPU compute needed for a request, and each model lists how its inputs map to Neurons.
Which models does Workers AI offer?
The catalog has more than 50 open-source models, including Llama, Mistral and DeepSeek families for text, plus models for embeddings, image classification, object detection, speech recognition, translation and text-to-image.
Can I pay with AI Gateway credits?
Yes. Set the gateway's Workers AI billing to Unified billing and use that gateway. Requests to frontier models paid with credits get higher rate limits.
Does Workers AI have an API?
Yes. You can invoke models from a Worker binding or through the Cloudflare REST API from any environment.
What are alternatives to Workers AI?
Alternatives for hosted open-source model inference include Replicate, Together AI, Fireworks AI, Amazon Bedrock and Hugging Face Inference Endpoints.
Sources & verification
Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.
Image credits
- Logo: from developers.cloudflare.com · fetched · Vendor trademark, used only to identify the product; removed on request
- Cover: none collected yet, so a generated tile is shown.
Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.
More in Developer platforms 4
All developer platforms toolsOpen source, Kubernetes-native AI platform made of modular projects for training, tuning, pipelines, notebooks and model management.
Open sourceNix flake that packages the OpenClaw agent gateway so macOS and headless Linux machines can run it from one declarative config.
Open sourceAutomated software testing suite for static analysis, API testing, Selenium UI tests and service virtualization, with AI-assisted test creation.
Contact sales
