cloudflare-workers-ai

Name: cloudflare-workers-ai
Author: jackspace

15stars7forks

How to install cloudflare-workers-ai?

jackspace/claudeskillz/cloudflare-workers-ai

$npx -y skills add jackspace/ClaudeSkillz --skill cloudflare-workers-ai

Installs into the current project.

Summary

Complete knowledge domain for Cloudflare Workers AI - Run AI models on serverless GPUs across Cloudflare's global network. Use when: implementing AI inference on Workers, running LLM models, generating text/images with AI, configuring Workers AI bindings, implementing AI streamin

Files · 1

View on GitHub

SKILL.md

1	# Cloudflare Workers AI - Complete Reference
2
3	Production-ready knowledge domain for building AI-powered applications with Cloudflare Workers AI.
4
5	Status: Production Ready ✅
6	Last Updated: 2025-10-21
7	Dependencies: cloudflare-worker-base (for Worker setup)
8	Latest Versions: wrangler@4.43.0, @cloudflare/workers-types@4.20251014.0
9
10	---
11
12	## Table of Contents
13
14	1. [Quick Start (5 minutes)](#quick-start-5-minutes)
15	2. [Workers AI API Reference](#workers-ai-api-reference)
16	3. [Model Selection Guide](#model-selection-guide)
17	4. [Common Patterns](#common-patterns)
18	5. [AI Gateway Integration](#ai-gateway-integration)
19	6. [Rate Limits & Pricing](#rate-limits--pricing)
20	7. [Production Checklist](#production-checklist)
21
22	---
23
24	## Quick Start (5 minutes)
25
26	### 1. Add AI Binding
27
28	wrangler.jsonc:
29	```jsonc
30	{
31	"ai": {
32	"binding": "AI"
33	}
34	}
35	```
36
37	### 2. Run Your First Model
38
39	```typescript
40	export interface Env {
41	AI: Ai;
42	}
43
44	export default {
45	async fetch(request: Request, env: Env): Promise<Response> {
46	const response = await env.AI.run('@cf/meta/llama-3.1-8b-instruct', {
47	prompt: 'What is Cloudflare?',
48	});
49
50	return Response.json(response);
51	},
52	};
53	```
54
55	### 3. Add Streaming (Recommended)
56
57	```typescript
58	const stream = await env.AI.run('@cf/meta/llama-3.1-8b-instruct', {
59	messages: [{ role: 'user', content: 'Tell me a story' }],
60	stream: true, // Always use streaming for text generation!
61	});
62
63	return new Response(stream, {
64	headers: { 'content-type': 'text/event-stream' },
65	});
66	```
67
68	Why streaming?
69	- Prevents buffering large responses in memory
70	- Faster time-to-first-token
71	- Better user experience for long-form content
72	- Avoids Worker timeout issues
73
74	---
75
76	## Workers AI API Reference
77
78	### `env.AI.run()`
79
80	Run an AI model inference.
81
82	Signature:
83	```typescript
84	async env.AI.run(
85	model: string,
86	inputs: ModelInputs,
87	options?: { gateway?: { id: string; skipCache?: boolean } }
88	): Promise<ModelOutput \| ReadableStream>
89	```
90
91	Parameters:
92
93	- `model` (string, required) - Model ID (e.g., `@cf/meta/llama-3.1-8b-instruct`)
94	- `inputs` (object, required) - Model-specific inputs
95	- `options` (object, optional) - Additional options
96	- `gateway` (object) - AI Gateway configuration
97	- `id` (string) - Gateway ID
98	- `skipCache` (boolean) - Skip AI Gateway cache
99
100	Returns:
101
102	- Non-streaming: `Promise<ModelOutput>` - JSON response
103	- Streaming: `ReadableStream` - Server-sent events stream
104
105	---
106
107	### Text Generation Models
108
109	Input Format:
110	```typescript
111	{
112	messages?: Array<{ role: 'system' \| 'user' \| 'assistant'; content: string }>;
113	prompt?: string; // Deprecated, use messages
114	stream?: boolean; // Default: false
115	max_tokens?: number; // Max tokens to generate
116	temperature?: number; // 0.0-1.0, default varies by model
117	top_p?: number; // 0.0-1.0
118	top_k?: number;
119	}
120	```
121
122	Output Format (Non-Streaming):
123	```typescript
124	{
125	response: string; // Generated text
126	}
127	```
128
129	Example:
130	```typescript
131	const response = await env.AI.run('@cf/meta/llama-3.1-8b-instruct', {
132	messages: [
133	{ role: 'system', content: 'You are a helpful assistant.' },
134	{ role: 'user', content: 'What is TypeScript?' },
135	],
136	stream: false,
137	});
138
139	console.log(response.response);
140	```
141
142	---
143
144	### Text Embeddings Models
145
146	Input Format:
147	```typescript
148	{
149	text: string \| string[]; // Single text or array of texts
150	}
151	```
152
153	Output Format:
154	```typescript
155	{
156	shape: number[]; // [batch_size, embedding_dimensions]
157	data: number[][]; // Array of embedding vectors
158	}
159	```
160
161	Example:
162	```typescript
163	const embeddings = await env.AI.run('@cf/baai/bge-base-en-v1.5', {
164	text: ['Hello world', 'Cloudflare Workers'],
165	});
166
167	console.log(embeddings.shape); // [2, 768]
168	console.log(embeddings.data[0]); // [0.123, -0.456, ...]
169	```
170
171	---
172
173	### Image Generation Models
174
175	Input Format:
176	```typescript
177	{
178	prompt: string; // Text description
179	num_steps?: number;

People who liked this also viewed

6 picks

Type

cloudflare-workers-ai

15stars7forks

View on GitHub

How to install cloudflare-workers-ai?

jackspace/claudeskillz/cloudflare-workers-ai

$npx -y skills add jackspace/ClaudeSkillz --skill cloudflare-workers-ai

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Cloudflare Workers AI - Complete Reference
2
3	Production-ready knowledge domain for building AI-powered applications with Cloudflare Workers AI.
4
5	Status: Production Ready ✅
6	Last Updated: 2025-10-21
7	Dependencies: cloudflare-worker-base (for Worker setup)
8	Latest Versions: wrangler@4.43.0, @cloudflare/workers-types@4.20251014.0
9
10	---
11
12	## Table of Contents
13
14	1. [Quick Start (5 minutes)](#quick-start-5-minutes)
15	2. [Workers AI API Reference](#workers-ai-api-reference)
16	3. [Model Selection Guide](#model-selection-guide)
17	4. [Common Patterns](#common-patterns)
18	5. [AI Gateway Integration](#ai-gateway-integration)
19	6. [Rate Limits & Pricing](#rate-limits--pricing)
20	7. [Production Checklist](#production-checklist)
21
22	---
23
24	## Quick Start (5 minutes)
25
26	### 1. Add AI Binding
27
28	wrangler.jsonc:
29	```jsonc
30	{
31	"ai": {
32	"binding": "AI"
33	}
34	}
35	```
36
37	### 2. Run Your First Model
38
39	```typescript
40	export interface Env {
41	AI: Ai;
42	}
43
44	export default {
45	async fetch(request: Request, env: Env): Promise<Response> {
46	const response = await env.AI.run('@cf/meta/llama-3.1-8b-instruct', {
47	prompt: 'What is Cloudflare?',
48	});
49
50	return Response.json(response);
51	},
52	};
53	```
54
55	### 3. Add Streaming (Recommended)
56
57	```typescript
58	const stream = await env.AI.run('@cf/meta/llama-3.1-8b-instruct', {
59	messages: [{ role: 'user', content: 'Tell me a story' }],
60	stream: true, // Always use streaming for text generation!
61	});
62
63	return new Response(stream, {
64	headers: { 'content-type': 'text/event-stream' },
65	});
66	```
67
68	Why streaming?
69	- Prevents buffering large responses in memory
70	- Faster time-to-first-token
71	- Better user experience for long-form content
72	- Avoids Worker timeout issues
73
74	---
75
76	## Workers AI API Reference
77
78	### `env.AI.run()`
79
80	Run an AI model inference.
81
82	Signature:
83	```typescript
84	async env.AI.run(
85	model: string,
86	inputs: ModelInputs,
87	options?: { gateway?: { id: string; skipCache?: boolean } }
88	): Promise<ModelOutput \| ReadableStream>
89	```
90
91	Parameters:
92
93	- `model` (string, required) - Model ID (e.g., `@cf/meta/llama-3.1-8b-instruct`)
94	- `inputs` (object, required) - Model-specific inputs
95	- `options` (object, optional) - Additional options
96	- `gateway` (object) - AI Gateway configuration
97	- `id` (string) - Gateway ID
98	- `skipCache` (boolean) - Skip AI Gateway cache
99
100	Returns:
101
102	- Non-streaming: `Promise<ModelOutput>` - JSON response
103	- Streaming: `ReadableStream` - Server-sent events stream
104
105	---
106
107	### Text Generation Models
108
109	Input Format:
110	```typescript
111	{
112	messages?: Array<{ role: 'system' \| 'user' \| 'assistant'; content: string }>;
113	prompt?: string; // Deprecated, use messages
114	stream?: boolean; // Default: false
115	max_tokens?: number; // Max tokens to generate
116	temperature?: number; // 0.0-1.0, default varies by model
117	top_p?: number; // 0.0-1.0
118	top_k?: number;
119	}
120	```
121
122	Output Format (Non-Streaming):
123	```typescript
124	{
125	response: string; // Generated text
126	}
127	```
128
129	Example:
130	```typescript
131	const response = await env.AI.run('@cf/meta/llama-3.1-8b-instruct', {
132	messages: [
133	{ role: 'system', content: 'You are a helpful assistant.' },
134	{ role: 'user', content: 'What is TypeScript?' },
135	],
136	stream: false,
137	});
138
139	console.log(response.response);
140	```
141
142	---
143
144	### Text Embeddings Models
145
146	Input Format:
147	```typescript
148	{
149	text: string \| string[]; // Single text or array of texts
150	}
151	```
152
153	Output Format:
154	```typescript
155	{
156	shape: number[]; // [batch_size, embedding_dimensions]
157	data: number[][]; // Array of embedding vectors
158	}
159	```
160
161	Example:
162	```typescript
163	const embeddings = await env.AI.run('@cf/baai/bge-base-en-v1.5', {
164	text: ['Hello world', 'Cloudflare Workers'],
165	});
166
167	console.log(embeddings.shape); // [2, 768]
168	console.log(embeddings.data[0]); // [0.123, -0.456, ...]
169	```
170
171	---
172
173	### Image Generation Models
174
175	Input Format:
176	```typescript
177	{
178	prompt: string; // Text description
179	num_steps?: number;

People who liked this also viewed

6 picks

Type