llmkit 0.1.3

Production-grade LLM client - 100+ providers, 11,000+ models. Pure Rust.
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

"""
Retry & Resilience Example

Demonstrates configuring automatic retry behavior for handling
transient failures like rate limits, timeouts, and server errors.

Requirements:
- Set OPENROUTER_API_KEY environment variable
- uv sync

Run:
    uv run python 18_retry_resilience.py
"""

from llmkit import LLMKitClient, Message, CompletionRequest, RetryConfig


def main():
    print("Retry & Resilience Configuration Example")
    print("=" * 50)

    # ========================================
    # Example 1: Default Retry (Recommended)
    # ========================================
    print("\n1. Default Retry Configuration")
    print("-" * 40)
    print("Settings: 10 retries, 1s-5min exponential backoff, jitter enabled")

    client = LLMKitClient.from_env()  # Uses default retry

    request = CompletionRequest(
        model="openrouter/qwen/qwen3-32b",
        messages=[Message.user("Say 'retry test passed' in exactly 3 words.")],
        max_tokens=50,
    ).without_thinking()

    response = client.complete(request)
    print(f"Response: {response.text_content().strip()}")

    # ========================================
    # Example 2: Production Retry Config
    # ========================================
    print("\n2. Production Retry Configuration")
    print("-" * 40)

    production_config = RetryConfig.production()
    print(f"Settings:")
    print(f"  Max retries: {production_config.max_retries}")
    print(f"  Initial delay: {production_config.initial_delay_ms}ms")
    print(f"  Max delay: {production_config.max_delay_ms}ms")
    print(f"  Backoff multiplier: {production_config.backoff_multiplier}")
    print(f"  Jitter: {production_config.jitter}")

    client = LLMKitClient.from_env(retry_config=production_config)

    request = CompletionRequest(
        model="openrouter/qwen/qwen3-32b",
        messages=[Message.user("Say 'production config works'")],
        max_tokens=50,
    ).without_thinking()

    response = client.complete(request)
    print(f"Response: {response.text_content().strip()}")

    # ========================================
    # Example 3: Conservative Retry Config
    # ========================================
    print("\n3. Conservative Retry Configuration")
    print("-" * 40)

    conservative_config = RetryConfig.conservative()
    print(f"Settings:")
    print(f"  Max retries: {conservative_config.max_retries}")
    print(f"  Initial delay: {conservative_config.initial_delay_ms}ms")
    print(f"  Max delay: {conservative_config.max_delay_ms}ms")
    print("Best for: Quick failures, interactive applications")

    client = LLMKitClient.from_env(retry_config=conservative_config)

    request = CompletionRequest(
        model="openrouter/qwen/qwen3-32b",
        messages=[Message.user("Say 'conservative works'")],
        max_tokens=50,
    ).without_thinking()

    response = client.complete(request)
    print(f"Response: {response.text_content().strip()}")

    # ========================================
    # Example 4: Custom Retry Config
    # ========================================
    print("\n4. Custom Retry Configuration")
    print("-" * 40)

    custom_config = RetryConfig(
        max_retries=5,
        initial_delay_ms=500,
        max_delay_ms=10000,
        backoff_multiplier=1.5,
        jitter=True,
    )

    print("Custom settings:")
    print(f"  Max retries: {custom_config.max_retries}")
    print(f"  Initial delay: {custom_config.initial_delay_ms}ms")
    print(f"  Max delay: {custom_config.max_delay_ms}ms")
    print(f"  Backoff multiplier: {custom_config.backoff_multiplier}")
    print(f"  Jitter: {custom_config.jitter}")

    client = LLMKitClient.from_env(retry_config=custom_config)

    request = CompletionRequest(
        model="openrouter/qwen/qwen3-32b",
        messages=[Message.user("Say 'custom config works'")],
        max_tokens=50,
    ).without_thinking()

    response = client.complete(request)
    print(f"Response: {response.text_content().strip()}")

    # ========================================
    # Example 5: No Retry (for testing/debugging)
    # ========================================
    print("\n5. No Retry Configuration")
    print("-" * 40)
    print("Useful for: Testing, debugging, fail-fast scenarios")

    # Option 1: Using RetryConfig.none()
    client = LLMKitClient.from_env(retry_config=RetryConfig.none())

    # Option 2: Using False (Python only)
    # client = LLMKitClient.from_env(retry_config=False)

    request = CompletionRequest(
        model="openrouter/qwen/qwen3-32b",
        messages=[Message.user("Say 'no retry works'")],
        max_tokens=50,
    ).without_thinking()

    response = client.complete(request)
    print(f"Response: {response.text_content().strip()}")

    # Summary
    print("\n" + "=" * 50)
    print("Summary: Retry Configurations")
    print("=" * 50)
    print("""
Retry handles these transient errors automatically:
  - Rate limits (429)
  - Server errors (5xx)
  - Timeouts
  - Connection errors

Configuration options:
  - Default (None)              -> 10 retries, 5min max
  - RetryConfig.production()    -> Same as default
  - RetryConfig.conservative()  -> 3 retries, 30s max
  - RetryConfig.none()          -> No retries
  - RetryConfig(...)            -> Custom settings
  - False                       -> No retries (Python only)
""")


if __name__ == "__main__":
    main()