feat(ai-gateway-provider): add embedding model support

[agcty]

Summary

Adds embedding model support to ai-gateway-provider, allowing users to route embedding requests through AI Gateway.

Changes

• Add AiGatewayEmbeddingModel class implementing EmbeddingModelV3
• Add embedding(), textEmbedding(), and textEmbeddingModel() methods to the provider factory
• Extract shared types and utilities to shared.ts to avoid circular imports
• Add comprehensive test coverage for embedding functionality

Usage

import { createAiGateway } from "ai-gateway-provider";
import { createOpenAI } from "@ai-sdk/openai";
import { embed } from "ai";

const gateway = createAiGateway({
  accountId: "your-account-id",
  gateway: "your-gateway",
  apiKey: "your-api-key",
});

const openai = createOpenAI({ apiKey: "..." });

const { embedding } = await embed({
  model: gateway.embedding(openai.textEmbeddingModel("text-embedding-3-small")),
  value: "Hello, world!",
});

Test plan

• All existing tests pass (13 tests)
• New embedding tests pass (6 tests)
• Type checking passes
• Changeset included

🤖 Generated with Claude Code

[changeset-bot]

🦋 Changeset detected

Latest commit: 744b7a3

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
ai-gateway-provider	Minor

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[agcty]

Testing Notes

I tested this implementation locally by routing Google embedding requests through AI Gateway. The routing works correctly - requests are intercepted, reformatted for the Universal Endpoint, and return successful responses.

Debug logs from local testing (not included in final code):

[ai-gateway-embedding] doEmbed called, models: 1
[ai-gateway-embedding] Processing model: google.generative-ai gemini-embedding-001
[ai-gateway-embedding] FETCH INTERCEPTED: https://generativelanguage.googleapis.com/v1beta/models/gemini-embedding-001:embedContent
[ai-gateway-embedding] Requests captured: 1
[ai-gateway-embedding] Request body: [
  {
    "endpoint": "v1beta/models/gemini-embedding-001:embedContent",
    "headers": {
      "content-type": "application/json",
      "x-goog-api-key": "..."
    },
    "provider": "google-ai-studio",
    "query": {
      "model": "models/gemini-embedding-001",
      "content": { "parts": [{ "text": "..." }] },
      "outputDimensionality": 1024
    }
  }
]
[ai-gateway-embedding] Sending to AI Gateway: https://gateway.ai.cloudflare.com/v1/{accountId}/test
[ai-gateway-embedding] AI Gateway response status: 200

Note: Embedding requests don't appear in the AI Gateway dashboard, though they route correctly and return 200 responses. This may require backend support to display embedding analytics alongside chat completions.