fix(ai-openrouter): Add support for google-gemini-v1 reasoning format

[subtleGradient]

Problem

Google Gemini models (gemini-2.0-flash-exp, gemini-pro-preview, etc.) fail with schema validation errors when used through @effect/ai-openrouter:

MalformedOutput: Expected "unknown" | "openai-responses-v1" | "anthropic-claude-v1", 
                 actual "google-gemini-v1"

The error occurs because Gemini models return reasoning_details with format: "google-gemini-v1", which is not recognized by the schema.

Solution

Add "google-gemini-v1" to the three reasoning format schemas:

• ReasoningDetailSummaryFormat
• ReasoningDetailEncryptedFormat
• ReasoningDetailTextFormat

This is a backward-compatible change (union type extension). Existing models continue to work unchanged.

Minimal Reproduction

import * as OpenRouterClient from "@effect/ai-openrouter/OpenRouterClient";
import * as OpenRouterLanguageModel from "@effect/ai-openrouter/OpenRouterLanguageModel";
import * as LanguageModel from "@effect/ai/LanguageModel";
import * as Prompt from "@effect/ai/Prompt";
import { FetchHttpClient } from "@effect/platform";
import * as Layer from "effect/Layer";
import * as Effect from "effect/Effect";
import * as Redacted from "effect/Redacted";

const program = Effect.gen(function* () {
  const prompt = Prompt.make([
    { role: "user" as const, content: "Say hello" }
  ]);
  
  const response = yield* LanguageModel.generateText({ prompt });
  return response.text;
});

const OpenRouterClientLayer = OpenRouterClient.layer({
  apiKey: Redacted.make(process.env.OPENROUTER_API_KEY!),
}).pipe(Layer.provide(FetchHttpClient.layer));

const OpenRouterModelLayer = OpenRouterLanguageModel.layer({
  model: "google/gemini-2.0-flash-exp",
}).pipe(Layer.provide(OpenRouterClientLayer));

// Before fix: MalformedOutput error
// After fix: Works correctly
await Effect.runPromise(
  program.pipe(Effect.provide(OpenRouterModelLayer))
);

Error before fix:

MalformedOutput: Expected "unknown" | "openai-responses-v1" | "anthropic-claude-v1",
                 actual "google-gemini-v1"

Testing

• ✅ Package builds successfully
• ✅ Schema now accepts google-gemini-v1 format
• ✅ Existing formats (OpenAI, Anthropic) still work
• ✅ No breaking changes

Changes

Modified:

• packages/ai/openrouter/src/Generated.ts - Added "google-gemini-v1" to format schemas

Added:

• .changeset/google-gemini-reasoning-format.md - Patch-level changeset

[changeset-bot]

🦋 Changeset detected

Latest commit: 071989e

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

This PR includes changesets to release 1 package

Name	Type
@effect/ai-openrouter	Patch

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

[Copilot]

Pull request overview

This PR fixes schema validation errors that occurred when using Google Gemini models (e.g., gemini-2.0-flash-exp) through OpenRouter. The issue was caused by Gemini models returning reasoning_details with format: "google-gemini-v1", which wasn't recognized by the existing schema.

Key Changes:

• Extended three reasoning format schemas to include "google-gemini-v1" as a valid literal value
• Added appropriate changeset documentation for the patch-level change

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
packages/ai/openrouter/src/Generated.ts	Added "google-gemini-v1" to all three reasoning format schemas (ReasoningDetailSummaryFormat, ReasoningDetailEncryptedFormat, ReasoningDetailTextFormat)
.changeset/google-gemini-reasoning-format.md	Added patch-level changeset documenting the support for Google Gemini v1 reasoning format

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[IMax153]

@subtleGradient - getting a CI error due to linting.

I think running a simple pnpm lint-fix should solve the problem.

[subtleGradient]

I should definitely not ping @IMax153 too frequently. But... here goes :P