qai-sdk 0.1.21

Universal Rust SDK for AI Providers
Documentation 
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
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224

//! # OpenAI-Compatible Provider Example
//!
//! Demonstrates using the OpenAI-Compatible provider to connect to any
//! third-party API endpoint that follows the OpenAI API format.
//! Examples include: Together AI, Groq, Fireworks, Perplexity, Mistral,
//! Anyscale, OpenRouter, local Ollama, and vLLM.

use futures::StreamExt;
use qai_sdk::openai_compatible::OpenAICompatibleProviderSettings;
use qai_sdk::types::EmbeddingOptions;
use qai_sdk::EmbeddingModel;
use qai_sdk::*;

use qai_sdk::LanguageModel;
#[tokio::main]
async fn main() -> Result<()> {
    dotenvy::dotenv().ok();

    // ===================================================================
    // 1. Together AI
    // ===================================================================
    println!("=== Together AI ===");
    let together = create_openai_compatible(OpenAICompatibleProviderSettings {
        base_url: "https://api.together.xyz/v1".to_string(),
        name: "together".to_string(),
        api_key: Some(std::env::var("TOGETHER_API_KEY").unwrap_or_default()),
        headers: None,
    });

    // Chat
    let model = together.chat("meta-llama/Llama-3-70b-chat-hf");
    let prompt = Prompt {
        messages: vec![Message {
            role: Role::User,
            content: vec![Content::Text {
                text: "What is Rust? One sentence.".to_string(),
            }],
        }],
    };
    let options = GenerateOptions {
        model_id: "meta-llama/Llama-3-70b-chat-hf".to_string(),
        max_tokens: Some(100),
        temperature: Some(0.7),
        top_p: None,
        stop_sequences: None,
        ..Default::default()
    };
    let result = model.generate(prompt.clone(), options.clone()).await?;
    println!("Together Chat: {}", result.text);

    // Embedding
    let model = together.embedding("togethercomputer/m2-bert-80M-8k-retrieval");
    let result = model
        .embed(
            vec!["Hello world".to_string()],
            EmbeddingOptions {
                model_id: "togethercomputer/m2-bert-80M-8k-retrieval".to_string(),
                dimensions: None,
            },
        )
        .await?;
    println!(
        "Together Embedding: {} dimensions\n",
        result.embeddings[0].len()
    );

    // ===================================================================
    // 2. Groq (Ultra-fast inference)
    // ===================================================================
    println!("=== Groq ===");
    let groq = create_openai_compatible(OpenAICompatibleProviderSettings {
        base_url: "https://api.groq.com/openai/v1".to_string(),
        name: "groq".to_string(),
        api_key: Some(std::env::var("GROQ_API_KEY").unwrap_or_default()),
        headers: None,
    });

    let model = groq.chat("llama-3.1-70b-versatile");
    let options = GenerateOptions {
        model_id: "llama-3.1-70b-versatile".to_string(),
        max_tokens: Some(100),
        temperature: Some(0.5),
        top_p: None,
        stop_sequences: None,
        ..Default::default()
    };
    let result = model.generate(prompt.clone(), options.clone()).await?;
    println!("Groq response: {}\n", result.text);

    // ===================================================================
    // 3. Fireworks AI
    // ===================================================================
    println!("=== Fireworks AI ===");
    let fireworks = create_openai_compatible(OpenAICompatibleProviderSettings {
        base_url: "https://api.fireworks.ai/inference/v1".to_string(),
        name: "fireworks".to_string(),
        api_key: Some(std::env::var("FIREWORKS_API_KEY").unwrap_or_default()),
        headers: None,
    });

    let model = fireworks.chat("accounts/fireworks/models/llama-v3p1-70b-instruct");
    let options = GenerateOptions {
        model_id: "accounts/fireworks/models/llama-v3p1-70b-instruct".to_string(),
        max_tokens: Some(100),
        temperature: None,
        top_p: None,
        stop_sequences: None,
        ..Default::default()
    };
    let result = model.generate(prompt.clone(), options.clone()).await?;
    println!("Fireworks response: {}\n", result.text);

    // ===================================================================
    // 4. OpenRouter (Multi-provider gateway)
    // ===================================================================
    println!("=== OpenRouter ===");
    let openrouter = create_openai_compatible(OpenAICompatibleProviderSettings {
        base_url: "https://openrouter.ai/api/v1".to_string(),
        name: "openrouter".to_string(),
        api_key: Some(std::env::var("OPENROUTER_API_KEY").unwrap_or_default()),
        headers: None,
    });

    let model = openrouter.chat("meta-llama/llama-3.1-8b-instruct:free");
    let options = GenerateOptions {
        model_id: "meta-llama/llama-3.1-8b-instruct:free".to_string(),
        max_tokens: Some(100),
        temperature: None,
        top_p: None,
        stop_sequences: None,
        ..Default::default()
    };
    let result = model.generate(prompt.clone(), options.clone()).await?;
    println!("OpenRouter response: {}\n", result.text);

    // ===================================================================
    // 5. Mistral AI
    // ===================================================================
    println!("=== Mistral AI ===");
    let mistral = create_openai_compatible(OpenAICompatibleProviderSettings {
        base_url: "https://api.mistral.ai/v1".to_string(),
        name: "mistral".to_string(),
        api_key: Some(std::env::var("MISTRAL_API_KEY").unwrap_or_default()),
        headers: None,
    });

    let model = mistral.chat("mistral-small-latest");
    let options = GenerateOptions {
        model_id: "mistral-small-latest".to_string(),
        max_tokens: Some(100),
        temperature: None,
        top_p: None,
        stop_sequences: None,
        ..Default::default()
    };
    let result = model.generate(prompt.clone(), options.clone()).await?;
    println!("Mistral response: {}\n", result.text);

    // ===================================================================
    // 6. Local Ollama
    // ===================================================================
    println!("=== Local Ollama ===");
    let ollama = create_openai_compatible(OpenAICompatibleProviderSettings {
        base_url: "http://localhost:11434/v1".to_string(),
        name: "ollama".to_string(),
        api_key: Some("ollama".to_string()), // Ollama doesn't require a key
        headers: None,
    });

    let model = ollama.chat("llama3.2");
    let options = GenerateOptions {
        model_id: "llama3.2".to_string(),
        max_tokens: Some(100),
        temperature: None,
        top_p: None,
        stop_sequences: None,
        ..Default::default()
    };

    // Ollama may not be running, so handle errors gracefully
    match model.generate(prompt.clone(), options.clone()).await {
        Ok(result) => println!("Ollama response: {}", result.text),
        Err(e) => println!("Ollama not available: {} (start with `ollama serve`)", e),
    }

    // ===================================================================
    // 7. Streaming with Compatible Provider
    // ===================================================================
    println!("\n=== Streaming with Compatible Provider ===");
    let model = groq.chat("llama-3.1-8b-instant");
    let options = GenerateOptions {
        model_id: "llama-3.1-8b-instant".to_string(),
        max_tokens: Some(100),
        temperature: None,
        top_p: None,
        stop_sequences: None,
        ..Default::default()
    };

    let mut stream = model.generate_stream(prompt.clone(), options).await?;
    print!("Streaming: ");
    while let Some(part) = stream.next().await {
        if let StreamPart::TextDelta { delta } = part {
            print!("{}", delta);
        }
    }
    println!("\n");

    // ===================================================================
    // Summary of supported endpoints
    // ===================================================================
    println!("--- Supported OpenAI-Compatible Endpoints ---");
    println!("Together AI:  api.together.xyz/v1");
    println!("Groq:         api.groq.com/openai/v1");
    println!("Fireworks:    api.fireworks.ai/inference/v1");
    println!("OpenRouter:   openrouter.ai/api/v1");
    println!("Mistral:      api.mistral.ai/v1");
    println!("Perplexity:   api.perplexity.ai");
    println!("Ollama:       localhost:11434/v1");
    println!("vLLM:         localhost:8000/v1");
    println!("LM Studio:    localhost:1234/v1");

    Ok(())
}