Openrouter Upgrade Migration

OpenRouter Upgrade & Migration

Current State

!npm list openai 2>/dev/null | head -5 !pip show openai 2>/dev/null | head -5

Overview

Migrating to OpenRouter from a direct provider API (OpenAI, Anthropic) is minimal: change base_url and api_key, add two headers. The OpenAI SDK works natively with OpenRouter. This skill covers migrating from direct APIs, switching between models, upgrading SDK versions, and running comparison tests.

Prerequisites

• An existing direct OpenAI or Anthropic integration to migrate — the Current State block above checks your installed openai SDK via npm list openai / pip show openai
• An OpenRouter API key (sk-or-v1-...) exported as OPENROUTER_API_KEY — see the openrouter-install-auth skill for setup
• Python 3.8+ or Node.js 18+ with the OpenAI SDK (Anthropic SDK users switch to the OpenAI SDK as part of the migration)
• The old provider key (OPENAI_API_KEY / ANTHROPIC_API_KEY) kept active during migration for comparison tests and quick rollback

Instructions

1. Confirm your installed SDK versions from the Current State output at the top of this skill.
2. Apply the 3-line change per Migration from Direct OpenAI, Migration from Direct Anthropic, or TypeScript Migration: swap base_url to https://openrouter.ai/api/v1, switch to OPENROUTER_API_KEY, and add the HTTP-Referer / X-Title headers. Anthropic migrations also change response parsing to .choices[0].message.content.
3. Prefix every model ID with its provider per the Model ID Migration Map (e.g. gpt-4o → openai/gpt-4o).
4. Work through the Migration Checklist — config, code, testing, and operations items — before flipping traffic.
5. Run the Comparison Test Script on your critical prompts (temperature=0) to compare content, tokens, and latency against the old backend.
6. Roll out gradually with the Feature Flag Migration pattern (USE_OPENROUTER env var plus get_model_id mapping), moving 10% → 50% → 100%.
7. Watch for post-migration failures (401, model_not_found, response-format drift, +50–100ms latency) per the Error Handling table.

Migration from Direct OpenAI

# BEFORE: Direct OpenAI
from openai import OpenAI
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello"}],
    max_tokens=200,
)

# AFTER: Via OpenRouter (3 lines changed)
from openai import OpenAI
client = OpenAI(
    base_url="https://openrouter.ai/api/v1",     # ← Changed
    api_key=os.environ["OPENROUTER_API_KEY"],     # ← Changed
    default_headers={                              # ← Added
        "HTTP-Referer": "https://my-app.com",
        "X-Title": "my-app",
    },
)
response = client.chat.completions.create(
    model="openai/gpt-4o",  # ← Add provider prefix
    messages=[{"role": "user", "content": "Hello"}],
    max_tokens=200,
)

Migration from Direct Anthropic

# BEFORE: Direct Anthropic SDK
import anthropic
client = anthropic.Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])
response = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=200,
    messages=[{"role": "user", "content": "Hello"}],
)
content = response.content[0].text

# AFTER: Via OpenRouter (using OpenAI SDK instead of Anthropic SDK)
from openai import OpenAI
client = OpenAI(
    base_url="https://openrouter.ai/api/v1",
    api_key=os.environ["OPENROUTER_API_KEY"],
    default_headers={
        "HTTP-Referer": "https://my-app.com",
        "X-Title": "my-app",
    },
)
response = client.chat.completions.create(
    model="anthropic/claude-3.5-sonnet",  # OpenRouter model ID
    messages=[{"role": "user", "content": "Hello"}],
    max_tokens=200,
)
content = response.choices[0].message.content  # OpenAI response format

TypeScript Migration

// BEFORE: Direct OpenAI
import OpenAI from "openai";
const client = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });

// AFTER: Via OpenRouter
const client = new OpenAI({
  baseURL: "https://openrouter.ai/api/v1",
  apiKey: process.env.OPENROUTER_API_KEY,
  defaultHeaders: {
    "HTTP-Referer": "https://my-app.com",
    "X-Title": "my-app",
  },
});
// Change model from "gpt-4o" to "openai/gpt-4o"

Migration Checklist

MIGRATION_CHECKLIST = {
    "config": [
        "base_url changed to https://openrouter.ai/api/v1",
        "API key changed to OPENROUTER_API_KEY (sk-or-v1-...)",
        "HTTP-Referer and X-Title headers added",
        "Model IDs prefixed with provider/ (e.g., openai/gpt-4o)",
    ],
    "code": [
        "All client initialization updated",
        "Model IDs updated in all routes/configs",
        "Error handling covers OpenRouter-specific codes (402, 408)",
        "Streaming still works with new endpoint",
        "Tool/function calling still works",
    ],
    "testing": [
        "Same prompts produce comparable quality output",
        "Latency within acceptable range (expect +50-100ms)",
        "Token counts match expectations",
        "Cost tracking updated for OpenRouter pricing",
        "Fallback chain tested",
    ],
    "operations": [
        "Credit balance sufficient for expected usage",
        "Per-key credit limits configured",
        "Monitoring updated to track OpenRouter metrics",
        "Alerting on new error codes (402, 408)",
        "Rollback plan documented",
    ],
}

Model ID Migration Map

Comparison Test Script

def compare_migration(prompt: str, old_model: str, new_model: str):
    """Run same prompt through old and new configurations to compare."""
    import time

    # New: OpenRouter
    or_client = OpenAI(
        base_url="https://openrouter.ai/api/v1",
        api_key=os.environ["OPENROUTER_API_KEY"],
        default_headers={"HTTP-Referer": "https://my-app.com", "X-Title": "migration-test"},
    )

    start = time.monotonic()
    or_response = or_client.chat.completions.create(
        model=new_model,
        messages=[{"role": "user", "content": prompt}],
        max_tokens=200, temperature=0,
    )
    or_latency = (time.monotonic() - start) * 1000

    return {
        "openrouter": {
            "model": or_response.model,
            "content": or_response.choices[0].message.content[:100],
            "tokens": or_response.usage.prompt_tokens + or_response.usage.completion_tokens,
            "latency_ms": round(or_latency),
        },
    }

# Test
result = compare_migration(
    "What is 2+2?",
    old_model="gpt-4o",
    new_model="openai/gpt-4o",
)
print(json.dumps(result, indent=2))

Feature Flag Migration

import os

USE_OPENROUTER = os.environ.get("USE_OPENROUTER", "false").lower() == "true"

def get_llm_client():
    """Feature flag for gradual migration."""
    if USE_OPENROUTER:
        return OpenAI(
            base_url="https://openrouter.ai/api/v1",
            api_key=os.environ["OPENROUTER_API_KEY"],
            default_headers={"HTTP-Referer": "https://my-app.com", "X-Title": "my-app"},
        )
    else:
        return OpenAI(api_key=os.environ["OPENAI_API_KEY"])

def get_model_id(model: str) -> str:
    """Map model IDs based on current backend."""
    if USE_OPENROUTER and "/" not in model:
        MODEL_MAP = {"gpt-4o": "openai/gpt-4o", "gpt-4o-mini": "openai/gpt-4o-mini"}
        return MODEL_MAP.get(model, f"openai/{model}")
    return model

Output

• Migrated client initialization code: 3 changed lines (base_url, api_key, headers) plus provider-prefixed model IDs across routes/configs
• A comparison test JSON per prompt with the served model, a content preview, combined token count, and latency_ms
• A four-category migration checklist (config / code / testing / operations) to track cutover readiness
• A feature-flagged get_llm_client() that flips between direct OpenAI and OpenRouter via the USE_OPENROUTER env var

Examples

Verify a migrated model on the same prompt before flipping traffic:

result = compare_migration("What is 2+2?", old_model="gpt-4o", new_model="openai/gpt-4o")
print(json.dumps(result, indent=2))
# {
#   "openrouter": {
#     "model": "openai/gpt-4o",
#     "content": "2 + 2 = 4",
#     "tokens": 21,
#     "latency_ms": 934
#   }
# }

Expect OpenRouter latency to run ~50-100ms above the direct API. More worked examples: references/examples.md.

Error Handling

Enterprise Considerations

• Migration from direct provider to OpenRouter requires only 3 lines of code change
• Use feature flags for gradual migration (10% -> 50% -> 100%)
• Run comparison tests on critical prompts before full migration
• OpenRouter adds ~50-100ms overhead; use streaming to mask perceived latency
• Keep direct provider keys active during migration for quick rollback
• Update monitoring dashboards for OpenRouter-specific metrics (generation_id, provider used)

References

• Examples | Errors
• Quickstart | OpenAI Compatibility

Source: jeremylongshore/claude-code-plugins-plus-skills → skills/.curated/openrouter-upgrade-migration/SKILL.md

Also appears in: jeremylongshore/claude-code-plugins-plus-skills/plugins/saas-packs/openrouter-pack/skills/openrouter-upgrade-migration/SKILL.md