Update ai package CHANGELOG.md for v0.30.2+ changes

Part of #378

packages/ai/CHANGELOG.md

@@ -4,7 +4,23 @@
 
 ### Breaking Changes
 
-- **Agent API moved**: All agent functionality (`agentLoop`, `agentLoopContinue`, `AgentContext`, `AgentEvent`, `AgentTool`, `AgentToolResult`, etc.) has moved to `@mariozechner/pi-agent-core`. See the [agent-core README](../agent/README.md) for documentation.
+- **Agent API moved**: All agent functionality (`agentLoop`, `agentLoopContinue`, `AgentContext`, `AgentEvent`, `AgentTool`, `AgentToolResult`, etc.) has moved to `@mariozechner/pi-agent-core`. Import from that package instead of `@mariozechner/pi-ai`.
+
+### Added
+
+- **`GoogleThinkingLevel` type**: Exported type that mirrors Google's `ThinkingLevel` enum values (`"THINKING_LEVEL_UNSPECIFIED" | "MINIMAL" | "LOW" | "MEDIUM" | "HIGH"`). Allows configuring Gemini thinking levels without importing from `@google/genai`.
+- **`ANTHROPIC_OAUTH_TOKEN` env var**: Now checked before `ANTHROPIC_API_KEY` in `getEnvApiKey()`, allowing OAuth tokens to take precedence.
+- **`event-stream.js` export**: `AssistantMessageEventStream` utility now exported from package index.
+
+### Changed
+
+- **OAuth uses Web Crypto API**: PKCE generation and OAuth flows now use Web Crypto API (`crypto.subtle`) instead of Node.js `crypto` module. This improves browser compatibility while still working in Node.js 20+.
+- **Deterministic model generation**: `generate-models.ts` now sorts providers and models alphabetically for consistent output across runs. ([#332](https://github.com/badlogic/pi-mono/pull/332) by [@mrexodia](https://github.com/mrexodia))
+
+### Fixed
+
+- **OpenAI completions empty content blocks**: Empty text or thinking blocks in assistant messages are now filtered out before sending to the OpenAI completions API, preventing validation errors. ([#344](https://github.com/badlogic/pi-mono/pull/344) by [@default-anton](https://github.com/default-anton))
+- **zAi provider API mapping**: Fixed zAi models to use `openai-completions` API with correct base URL (`https://api.z.ai/api/coding/paas/v4`) instead of incorrect Anthropic API mapping. ([#344](https://github.com/badlogic/pi-mono/pull/344), [#358](https://github.com/badlogic/pi-mono/pull/358) by [@default-anton](https://github.com/default-anton))
 
 ## [0.28.0] - 2025-12-25
 

packages/ai/scripts/generate-models.ts

@@ -260,32 +260,32 @@ async function loadModelsDevData(): Promise<Model<any>[]> {
 
 		// Process zAi models
 		if (data.zai?.models) {
-      for (const [modelId, model] of Object.entries(data.zai.models)) {
+			for (const [modelId, model] of Object.entries(data.zai.models)) {
-        const m = model as ModelsDevModel;
+				const m = model as ModelsDevModel;
-        if (m.tool_call !== true) continue;
+				if (m.tool_call !== true) continue;
-        const supportsImage = m.modalities?.input?.includes("image")
+				const supportsImage = m.modalities?.input?.includes("image")
 
-        models.push({
+				models.push({
-          id: modelId,
+				id: modelId,
-          name: m.name || modelId,
+				name: m.name || modelId,
-          api: "openai-completions",
+				api: "openai-completions",
-          provider: "zai",
+				provider: "zai",
-          baseUrl: "https://api.z.ai/api/coding/paas/v4",
+				baseUrl: "https://api.z.ai/api/coding/paas/v4",
-          reasoning: m.reasoning === true,
+				reasoning: m.reasoning === true,
-          input: supportsImage ? ["text", "image"] : ["text"],
+				input: supportsImage ? ["text", "image"] : ["text"],
-          cost: {
+				cost: {
-            input: m.cost?.input || 0,
+					input: m.cost?.input || 0,
-            output: m.cost?.output || 0,
+					output: m.cost?.output || 0,
-            cacheRead: m.cost?.cache_read || 0,
+					cacheRead: m.cost?.cache_read || 0,
-            cacheWrite: m.cost?.cache_write || 0,
+					cacheWrite: m.cost?.cache_write || 0,
-          },
+				},
-          compat: {
+				compat: {
-            supportsDeveloperRole: false,
+					supportsDeveloperRole: false,
-          },
+				},
-          contextWindow: m.limit?.context || 4096,
+				contextWindow: m.limit?.context || 4096,
-          maxTokens: m.limit?.output || 4096,
+				maxTokens: m.limit?.output || 4096,
-        });
+				});
-      }
+			}
 		}
 
 		// Process Mistral models