ai-sdk-core

Name: ai-sdk-core
Author: jackspace

15stars7forks

How to install ai-sdk-core?

jackspace/claudeskillz/ai-sdk-core

$npx -y skills add jackspace/ClaudeSkillz --skill ai-sdk-core

Installs into the current project.

Summary

Backend AI functionality with Vercel AI SDK v5 - text generation, structured output with Zod, tool calling, and agents. Multi-provider support for OpenAI, Anthropic, Google, and Cloudflare Workers AI. Use when: implementing server-side AI features, generating text/chat completion

Files · 1

View on GitHub

SKILL.md

1	# AI SDK Core
2
3	Production-ready backend AI with Vercel AI SDK v5.
4
5	## Quick Start (5 Minutes)
6
7	### Installation
8
9	```bash
10	# Core package
11	npm install ai
12
13	# Provider packages (install what you need)
14	npm install @ai-sdk/openai # OpenAI (GPT-5, GPT-4, GPT-3.5)
15	npm install @ai-sdk/anthropic # Anthropic (Claude Sonnet 4.5, Opus 4, Haiku 4)
16	npm install @ai-sdk/google # Google (Gemini 2.5 Pro/Flash/Lite)
17	npm install workers-ai-provider # Cloudflare Workers AI
18
19	# Schema validation
20	npm install zod
21	```
22
23	### Environment Variables
24
25	```bash
26	# .env
27	OPENAI_API_KEY=sk-...
28	ANTHROPIC_API_KEY=sk-ant-...
29	GOOGLE_GENERATIVE_AI_API_KEY=...
30	```
31
32	### First Example: Generate Text
33
34	```typescript
35	import { generateText } from 'ai';
36	import { openai } from '@ai-sdk/openai';
37
38	const result = await generateText({
39	model: openai('gpt-4-turbo'),
40	prompt: 'What is TypeScript?',
41	});
42
43	console.log(result.text);
44	```
45
46	### First Example: Streaming Chat
47
48	```typescript
49	import { streamText } from 'ai';
50	import { anthropic } from '@ai-sdk/anthropic';
51
52	const stream = streamText({
53	model: anthropic('claude-sonnet-4-5-20250929'),
54	messages: [
55	{ role: 'user', content: 'Tell me a story' },
56	],
57	});
58
59	for await (const chunk of stream.textStream) {
60	process.stdout.write(chunk);
61	}
62	```
63
64	### First Example: Structured Output
65
66	```typescript
67	import { generateObject } from 'ai';
68	import { openai } from '@ai-sdk/openai';
69	import { z } from 'zod';
70
71	const result = await generateObject({
72	model: openai('gpt-4'),
73	schema: z.object({
74	name: z.string(),
75	age: z.number(),
76	skills: z.array(z.string()),
77	}),
78	prompt: 'Generate a person profile for a software engineer',
79	});
80
81	console.log(result.object);
82	// { name: "Alice", age: 28, skills: ["TypeScript", "React"] }
83	```
84
85	---
86
87	## Core Functions
88
89	### generateText()
90
91	Generate text completion with optional tools and multi-step execution.
92
93	Signature:
94
95	```typescript
96	async function generateText(options: {
97	model: LanguageModel;
98	prompt?: string;
99	messages?: Array<ModelMessage>;
100	system?: string;
101	tools?: Record<string, Tool>;
102	maxOutputTokens?: number;
103	temperature?: number;
104	stopWhen?: StopCondition;
105	// ... other options
106	}): Promise<GenerateTextResult>
107	```
108
109	Basic Usage:
110
111	```typescript
112	import { generateText } from 'ai';
113	import { openai } from '@ai-sdk/openai';
114
115	const result = await generateText({
116	model: openai('gpt-4-turbo'),
117	prompt: 'Explain quantum computing',
118	maxOutputTokens: 500,
119	temperature: 0.7,
120	});
121
122	console.log(result.text);
123	console.log(`Tokens: ${result.usage.totalTokens}`);
124	```
125
126	With Messages (Chat Format):
127
128	```typescript
129	const result = await generateText({
130	model: openai('gpt-4-turbo'),
131	messages: [
132	{ role: 'system', content: 'You are a helpful assistant.' },
133	{ role: 'user', content: 'What is the weather?' },
134	{ role: 'assistant', content: 'I need your location.' },
135	{ role: 'user', content: 'San Francisco' },
136	],
137	});
138	```
139
140	With Tools:
141
142	```typescript
143	import { tool } from 'ai';
144	import { z } from 'zod';
145
146	const result = await generateText({
147	model: openai('gpt-4'),
148	tools: {
149	weather: tool({
150	description: 'Get the weather for a location',
151	inputSchema: z.object({
152	location: z.string(),
153	}),
154	execute: async ({ location }) => {
155	// API call here
156	return { temperature: 72, condition: 'sunny' };
157	},
158	}),
159	},
160	prompt: 'What is the weather in Tokyo?',
161	});
162	```
163
164	When to Use:
165	- Need final response (not streaming)
166	- Want to wait for tool executions to complete
167	- Simpler code when streaming not needed
168	- Building batch/scheduled tasks
169
170	Error Handling:
171
172	```typescript
173	import { AI_APICallError, AI_NoContentGeneratedError } from 'ai';
174
175	try {
176	const result = await generateText({
177	model: openai('gpt-4-turbo'),
178	prompt: 'Hello',
179	});
180	console.log(result.text);
181	} catch (error) {
182	if (error instanceof AI_APICallError) {
183	console.error('API call failed:', error.message);
184	// Check rate

People who liked this also viewed

6 picks

Type

ai-sdk-core

15stars7forks

View on GitHub

How to install ai-sdk-core?

jackspace/claudeskillz/ai-sdk-core

$npx -y skills add jackspace/ClaudeSkillz --skill ai-sdk-core

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# AI SDK Core
2
3	Production-ready backend AI with Vercel AI SDK v5.
4
5	## Quick Start (5 Minutes)
6
7	### Installation
8
9	```bash
10	# Core package
11	npm install ai
12
13	# Provider packages (install what you need)
14	npm install @ai-sdk/openai # OpenAI (GPT-5, GPT-4, GPT-3.5)
15	npm install @ai-sdk/anthropic # Anthropic (Claude Sonnet 4.5, Opus 4, Haiku 4)
16	npm install @ai-sdk/google # Google (Gemini 2.5 Pro/Flash/Lite)
17	npm install workers-ai-provider # Cloudflare Workers AI
18
19	# Schema validation
20	npm install zod
21	```
22
23	### Environment Variables
24
25	```bash
26	# .env
27	OPENAI_API_KEY=sk-...
28	ANTHROPIC_API_KEY=sk-ant-...
29	GOOGLE_GENERATIVE_AI_API_KEY=...
30	```
31
32	### First Example: Generate Text
33
34	```typescript
35	import { generateText } from 'ai';
36	import { openai } from '@ai-sdk/openai';
37
38	const result = await generateText({
39	model: openai('gpt-4-turbo'),
40	prompt: 'What is TypeScript?',
41	});
42
43	console.log(result.text);
44	```
45
46	### First Example: Streaming Chat
47
48	```typescript
49	import { streamText } from 'ai';
50	import { anthropic } from '@ai-sdk/anthropic';
51
52	const stream = streamText({
53	model: anthropic('claude-sonnet-4-5-20250929'),
54	messages: [
55	{ role: 'user', content: 'Tell me a story' },
56	],
57	});
58
59	for await (const chunk of stream.textStream) {
60	process.stdout.write(chunk);
61	}
62	```
63
64	### First Example: Structured Output
65
66	```typescript
67	import { generateObject } from 'ai';
68	import { openai } from '@ai-sdk/openai';
69	import { z } from 'zod';
70
71	const result = await generateObject({
72	model: openai('gpt-4'),
73	schema: z.object({
74	name: z.string(),
75	age: z.number(),
76	skills: z.array(z.string()),
77	}),
78	prompt: 'Generate a person profile for a software engineer',
79	});
80
81	console.log(result.object);
82	// { name: "Alice", age: 28, skills: ["TypeScript", "React"] }
83	```
84
85	---
86
87	## Core Functions
88
89	### generateText()
90
91	Generate text completion with optional tools and multi-step execution.
92
93	Signature:
94
95	```typescript
96	async function generateText(options: {
97	model: LanguageModel;
98	prompt?: string;
99	messages?: Array<ModelMessage>;
100	system?: string;
101	tools?: Record<string, Tool>;
102	maxOutputTokens?: number;
103	temperature?: number;
104	stopWhen?: StopCondition;
105	// ... other options
106	}): Promise<GenerateTextResult>
107	```
108
109	Basic Usage:
110
111	```typescript
112	import { generateText } from 'ai';
113	import { openai } from '@ai-sdk/openai';
114
115	const result = await generateText({
116	model: openai('gpt-4-turbo'),
117	prompt: 'Explain quantum computing',
118	maxOutputTokens: 500,
119	temperature: 0.7,
120	});
121
122	console.log(result.text);
123	console.log(`Tokens: ${result.usage.totalTokens}`);
124	```
125
126	With Messages (Chat Format):
127
128	```typescript
129	const result = await generateText({
130	model: openai('gpt-4-turbo'),
131	messages: [
132	{ role: 'system', content: 'You are a helpful assistant.' },
133	{ role: 'user', content: 'What is the weather?' },
134	{ role: 'assistant', content: 'I need your location.' },
135	{ role: 'user', content: 'San Francisco' },
136	],
137	});
138	```
139
140	With Tools:
141
142	```typescript
143	import { tool } from 'ai';
144	import { z } from 'zod';
145
146	const result = await generateText({
147	model: openai('gpt-4'),
148	tools: {
149	weather: tool({
150	description: 'Get the weather for a location',
151	inputSchema: z.object({
152	location: z.string(),
153	}),
154	execute: async ({ location }) => {
155	// API call here
156	return { temperature: 72, condition: 'sunny' };
157	},
158	}),
159	},
160	prompt: 'What is the weather in Tokyo?',
161	});
162	```
163
164	When to Use:
165	- Need final response (not streaming)
166	- Want to wait for tool executions to complete
167	- Simpler code when streaming not needed
168	- Building batch/scheduled tasks
169
170	Error Handling:
171
172	```typescript
173	import { AI_APICallError, AI_NoContentGeneratedError } from 'ai';
174
175	try {
176	const result = await generateText({
177	model: openai('gpt-4-turbo'),
178	prompt: 'Hello',
179	});
180	console.log(result.text);
181	} catch (error) {
182	if (error instanceof AI_APICallError) {
183	console.error('API call failed:', error.message);
184	// Check rate

People who liked this also viewed

6 picks

Type