ai-sdk-provider-codex-cli 0.7.0 → 1.0.0-beta.1

package/README.md

@@ -4,23 +4,34 @@
 [![npm downloads](https://img.shields.io/npm/dm/ai-sdk-provider-codex-cli.svg)](https://www.npmjs.com/package/ai-sdk-provider-codex-cli)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
 ![Node >= 18](https://img.shields.io/badge/node-%3E%3D18-43853d?logo=node.js&logoColor=white)
-![AI SDK v5](https://img.shields.io/badge/AI%20SDK-v5-000?logo=vercel&logoColor=white)
+![AI SDK v6 beta](https://img.shields.io/badge/AI%20SDK-v6%20beta-000?logo=vercel&logoColor=white)
 ![Modules: ESM + CJS](https://img.shields.io/badge/modules-ESM%20%2B%20CJS-3178c6)
 ![TypeScript](https://img.shields.io/badge/TypeScript-blue)
 [![PRs welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg)](https://github.com/ben-vargas/ai-sdk-provider-codex-cli/issues)
 [![Latest Release](https://img.shields.io/github/v/release/ben-vargas/ai-sdk-provider-codex-cli?display_name=tag)](https://github.com/ben-vargas/ai-sdk-provider-codex-cli/releases/latest)
 
-A community provider for Vercel AI SDK v5 that uses OpenAI’s Codex CLI (non‑interactive `codex exec`) to talk to GPT‑5.1 class models (`gpt-5.1`, the Codex-specific `gpt-5.1-codex`, the flagship `gpt-5.1-codex-max`, and the lightweight `gpt-5.1-codex-mini` slugs) with your ChatGPT Plus/Pro subscription. The provider spawns the Codex CLI process, parses its JSONL output, and adapts it to the AI SDK LanguageModelV2 interface. Legacy GPT-5 / GPT-5-codex slugs remain compatible for existing workflows.
+A community provider for Vercel AI SDK v6 (beta) that uses OpenAI's Codex CLI (non‑interactive `codex exec`) to talk to GPT‑5.1 class models (`gpt-5.1`, the Codex-specific `gpt-5.1-codex`, the flagship `gpt-5.1-codex-max`, and the lightweight `gpt-5.1-codex-mini` slugs) with your ChatGPT Plus/Pro subscription. The provider spawns the Codex CLI process, parses its JSONL output, and adapts it to the AI SDK LanguageModelV3 interface. Legacy GPT-5 / GPT-5-codex slugs remain compatible for existing workflows.
+
+> **⚠️ AI SDK v6 Beta**: This version (`1.0.0-beta.x`) requires AI SDK v6 beta packages. For AI SDK v5, use the `0.x` releases from the main branch.
 
 - Works with `generateText`, `streamText`, and `generateObject` (native JSON Schema support via `--output-schema`)
 - Uses ChatGPT OAuth from `codex login` (tokens in `~/.codex/auth.json`) or `OPENAI_API_KEY`
 - Node-only (spawns a local process); supports CI and local dev
+- **v1.0.0-beta.1**: AI SDK v6 migration with LanguageModelV3 interface
 - **v0.5.0**: Adds comprehensive logging system with verbose mode and custom logger support
 - **v0.3.0**: Adds comprehensive tool streaming support for monitoring autonomous tool execution
-- **v0.2.0 Breaking Changes**: Switched to `--experimental-json` and native schema enforcement (see [CHANGELOG](CHANGELOG.md))
+
+## Version Compatibility
+
+| Provider Version | AI SDK Version | NPM Tag   | Status | Branch       |
+| ---------------- | -------------- | --------- | ------ | ------------ |
+| 1.x.x-beta       | v6 (beta)      | `beta`    | Beta   | `ai-sdk-v6`  |
+| 0.x.x            | v5             | `latest`  | Stable | `main`       |
 
 ## Installation
 
+### For AI SDK v6 (beta)
+
 1. Install and authenticate Codex CLI
 
 ```bash
@@ -28,18 +39,24 @@ npm i -g @openai/codex
 codex login   # or set OPENAI_API_KEY
 ```
 
-> **⚠️ Version Requirement**: Requires Codex CLI **>= 0.42.0** for `--experimental-json` and `--output-schema` support. **>= 0.60.0 recommended** for `gpt-5.1-codex-max` and `xhigh` reasoning effort. If you supply your own Codex CLI (global install or custom `codexPath`/`allowNpx`), check it with `codex --version` and upgrade if needed. The optional dependency `@openai/codex` in this package pulls a compatible version automatically.
->
-> ```bash
-> npm i -g @openai/codex@latest
-> ```
+2. Install provider and AI SDK v6 beta
 
-2. Install provider and AI SDK
+```bash
+npm i ai@beta ai-sdk-provider-codex-cli@beta
+```
+
+### For AI SDK v5 (stable)
 
 ```bash
 npm i ai ai-sdk-provider-codex-cli
 ```
 
+> **⚠️ Codex CLI Version**: Requires Codex CLI **>= 0.42.0** for `--experimental-json` and `--output-schema` support. **>= 0.60.0 recommended** for `gpt-5.1-codex-max` and `xhigh` reasoning effort. If you supply your own Codex CLI (global install or custom `codexPath`/`allowNpx`), check it with `codex --version` and upgrade if needed. The optional dependency `@openai/codex` in this package pulls a compatible version automatically.
+>
+> ```bash
+> npm i -g @openai/codex@latest
+> ```
+
 ## Quick Start
 
 Text generation
@@ -95,7 +112,7 @@ console.log(object);
 
 ## Features
 
-- AI SDK v5 compatible (LanguageModelV2)
+- AI SDK v6 compatible (LanguageModelV3)
 - Streaming and non‑streaming
 - **Configurable logging** (v0.5.0+) - Verbose mode, custom loggers, or silent operation
 - **Tool streaming support** (v0.3.0+) - Monitor autonomous tool execution in real-time

package/dist/index.cjs

@@ -388,7 +388,7 @@ function resolveCodexPath(explicitPath, allowNpx) {
   }
 }
 var CodexCliLanguageModel = class {
-  specificationVersion = "v2";
+  specificationVersion = "v3";
   provider = "codex-cli";
   defaultObjectGenerationMode = "json";
   supportsImageUrls = false;
@@ -727,8 +727,8 @@ var CodexCliLanguageModel = class {
     const add = (setting, name) => {
       if (setting !== void 0)
         unsupported.push({
-          type: "unsupported-setting",
-          setting: name,
+          type: "unsupported",
+          feature: name,
           details: `Codex CLI does not support ${name}; it will be ignored.`
         });
     };
@@ -751,15 +751,21 @@ var CodexCliLanguageModel = class {
   extractUsage(evt) {
     const reported = evt.usage;
     if (!reported) return void 0;
-    const inputTokens = reported.input_tokens ?? 0;
-    const outputTokens = reported.output_tokens ?? 0;
+    const inputTotal = reported.input_tokens ?? 0;
+    const outputTotal = reported.output_tokens ?? 0;
     const cachedInputTokens = reported.cached_input_tokens ?? 0;
     return {
-      inputTokens,
-      outputTokens,
-      // totalTokens should not double-count cached tokens; track cached separately
-      totalTokens: inputTokens + outputTokens,
-      cachedInputTokens
+      inputTokens: {
+        total: inputTotal,
+        noCache: inputTotal - cachedInputTokens,
+        cacheRead: cachedInputTokens,
+        cacheWrite: void 0
+      },
+      outputTokens: {
+        total: outputTotal,
+        text: outputTotal,
+        reasoning: void 0
+      }
     };
   }
   getToolName(item) {
@@ -957,7 +963,10 @@ var CodexCliLanguageModel = class {
       `[codex-cli] Executing Codex CLI: ${cmd} with ${args.length} arguments, cwd: ${cwd ?? "default"}`
     );
     let text = "";
-    const usage = { inputTokens: 0, outputTokens: 0, totalTokens: 0 };
+    let usage = {
+      inputTokens: { total: 0, noCache: 0, cacheRead: 0, cacheWrite: void 0 },
+      outputTokens: { total: 0, text: 0, reasoning: void 0 }
+    };
     const finishReason = "stop";
     const startTime = Date.now();
     const child = child_process.spawn(cmd, args, { env, cwd, stdio: ["ignore", "pipe", "pipe"] });
@@ -1003,9 +1012,7 @@ var CodexCliLanguageModel = class {
             if (event.type === "turn.completed") {
               const usageEvent = this.extractUsage(event);
               if (usageEvent) {
-                usage.inputTokens = usageEvent.inputTokens;
-                usage.outputTokens = usageEvent.outputTokens;
-                usage.totalTokens = usageEvent.totalTokens;
+                usage = usageEvent;
               }
             }
             if (event.type === "item.completed" && this.getItemType(event.item) === "assistant_message" && typeof event.item?.text === "string") {
@@ -1040,11 +1047,12 @@ var CodexCliLanguageModel = class {
               );
               return;
             }
+            const totalTokens = (usage.inputTokens.total ?? 0) + (usage.outputTokens.total ?? 0);
             this.logger.info(
-              `[codex-cli] Request completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${usage.totalTokens}`
+              `[codex-cli] Request completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${totalTokens}`
             );
             this.logger.debug(
-              `[codex-cli] Token usage - Input: ${usage.inputTokens}, Output: ${usage.outputTokens}, Total: ${usage.totalTokens}`
+              `[codex-cli] Token usage - Input: ${usage.inputTokens.total ?? 0}, Output: ${usage.outputTokens.total ?? 0}, Total: ${totalTokens}`
             );
             resolve();
           } else {
@@ -1272,12 +1280,16 @@ var CodexCliLanguageModel = class {
             controller.enqueue({ type: "text-delta", id: textId, delta: finalText });
             controller.enqueue({ type: "text-end", id: textId });
           }
-          const usageSummary = lastUsage ?? { inputTokens: 0, outputTokens: 0, totalTokens: 0 };
+          const usageSummary = lastUsage ?? {
+            inputTokens: { total: 0, noCache: 0, cacheRead: 0, cacheWrite: void 0 },
+            outputTokens: { total: 0, text: 0, reasoning: void 0 }
+          };
+          const totalTokens = (usageSummary.inputTokens.total ?? 0) + (usageSummary.outputTokens.total ?? 0);
           this.logger.info(
-            `[codex-cli] Stream completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${usageSummary.totalTokens}`
+            `[codex-cli] Stream completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${totalTokens}`
           );
           this.logger.debug(
-            `[codex-cli] Token usage - Input: ${usageSummary.inputTokens}, Output: ${usageSummary.outputTokens}, Total: ${usageSummary.totalTokens}`
+            `[codex-cli] Token usage - Input: ${usageSummary.inputTokens.total ?? 0}, Output: ${usageSummary.outputTokens.total ?? 0}, Total: ${totalTokens}`
           );
           controller.enqueue({
             type: "finish",
@@ -1381,8 +1393,8 @@ function createCodexCli(options = {}) {
   };
   provider$1.languageModel = createModel;
   provider$1.chat = createModel;
-  provider$1.textEmbeddingModel = ((modelId) => {
-    throw new provider.NoSuchModelError({ modelId, modelType: "textEmbeddingModel" });
+  provider$1.embeddingModel = ((modelId) => {
+    throw new provider.NoSuchModelError({ modelId, modelType: "embeddingModel" });
   });
   provider$1.imageModel = ((modelId) => {
     throw new provider.NoSuchModelError({ modelId, modelType: "imageModel" });

package/dist/index.d.cts

@@ -1,4 +1,4 @@
-import { ProviderV2, LanguageModelV2 } from '@ai-sdk/provider';
+import { ProviderV3, LanguageModelV3 } from '@ai-sdk/provider';
 
 /**
  * Logger interface for custom logging.
@@ -251,11 +251,11 @@ interface CodexCliProviderOptions {
     rmcpClient?: boolean;
 }
 
-interface CodexCliProvider extends ProviderV2 {
-    (modelId: string, settings?: CodexCliSettings): LanguageModelV2;
-    languageModel(modelId: string, settings?: CodexCliSettings): LanguageModelV2;
-    chat(modelId: string, settings?: CodexCliSettings): LanguageModelV2;
-    textEmbeddingModel(modelId: string): never;
+interface CodexCliProvider extends ProviderV3 {
+    (modelId: string, settings?: CodexCliSettings): LanguageModelV3;
+    languageModel(modelId: string, settings?: CodexCliSettings): LanguageModelV3;
+    chat(modelId: string, settings?: CodexCliSettings): LanguageModelV3;
+    embeddingModel(modelId: string): never;
     imageModel(modelId: string): never;
 }
 declare function createCodexCli(options?: CodexCliProviderSettings): CodexCliProvider;
@@ -265,8 +265,8 @@ interface CodexLanguageModelOptions {
     id: string;
     settings?: CodexCliSettings;
 }
-declare class CodexCliLanguageModel implements LanguageModelV2 {
-    readonly specificationVersion: "v2";
+declare class CodexCliLanguageModel implements LanguageModelV3 {
+    readonly specificationVersion: "v3";
     readonly provider = "codex-cli";
     readonly defaultObjectGenerationMode: "json";
     readonly supportsImageUrls = false;
@@ -301,8 +301,8 @@ declare class CodexCliLanguageModel implements LanguageModelV2 {
     private emitToolInvocation;
     private emitToolResult;
     private handleSpawnError;
-    doGenerate(options: Parameters<LanguageModelV2['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV2['doGenerate']>>>;
-    doStream(options: Parameters<LanguageModelV2['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV2['doStream']>>>;
+    doGenerate(options: Parameters<LanguageModelV3['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV3['doGenerate']>>>;
+    doStream(options: Parameters<LanguageModelV3['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV3['doStream']>>>;
 }
 
 declare function isAuthenticationError(err: unknown): boolean;

package/dist/index.d.ts

@@ -1,4 +1,4 @@
-import { ProviderV2, LanguageModelV2 } from '@ai-sdk/provider';
+import { ProviderV3, LanguageModelV3 } from '@ai-sdk/provider';
 
 /**
  * Logger interface for custom logging.
@@ -251,11 +251,11 @@ interface CodexCliProviderOptions {
     rmcpClient?: boolean;
 }
 
-interface CodexCliProvider extends ProviderV2 {
-    (modelId: string, settings?: CodexCliSettings): LanguageModelV2;
-    languageModel(modelId: string, settings?: CodexCliSettings): LanguageModelV2;
-    chat(modelId: string, settings?: CodexCliSettings): LanguageModelV2;
-    textEmbeddingModel(modelId: string): never;
+interface CodexCliProvider extends ProviderV3 {
+    (modelId: string, settings?: CodexCliSettings): LanguageModelV3;
+    languageModel(modelId: string, settings?: CodexCliSettings): LanguageModelV3;
+    chat(modelId: string, settings?: CodexCliSettings): LanguageModelV3;
+    embeddingModel(modelId: string): never;
     imageModel(modelId: string): never;
 }
 declare function createCodexCli(options?: CodexCliProviderSettings): CodexCliProvider;
@@ -265,8 +265,8 @@ interface CodexLanguageModelOptions {
     id: string;
     settings?: CodexCliSettings;
 }
-declare class CodexCliLanguageModel implements LanguageModelV2 {
-    readonly specificationVersion: "v2";
+declare class CodexCliLanguageModel implements LanguageModelV3 {
+    readonly specificationVersion: "v3";
     readonly provider = "codex-cli";
     readonly defaultObjectGenerationMode: "json";
     readonly supportsImageUrls = false;
@@ -301,8 +301,8 @@ declare class CodexCliLanguageModel implements LanguageModelV2 {
     private emitToolInvocation;
     private emitToolResult;
     private handleSpawnError;
-    doGenerate(options: Parameters<LanguageModelV2['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV2['doGenerate']>>>;
-    doStream(options: Parameters<LanguageModelV2['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV2['doStream']>>>;
+    doGenerate(options: Parameters<LanguageModelV3['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV3['doGenerate']>>>;
+    doStream(options: Parameters<LanguageModelV3['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV3['doStream']>>>;
 }
 
 declare function isAuthenticationError(err: unknown): boolean;

package/dist/index.js

@@ -385,7 +385,7 @@ function resolveCodexPath(explicitPath, allowNpx) {
   }
 }
 var CodexCliLanguageModel = class {
-  specificationVersion = "v2";
+  specificationVersion = "v3";
   provider = "codex-cli";
   defaultObjectGenerationMode = "json";
   supportsImageUrls = false;
@@ -724,8 +724,8 @@ var CodexCliLanguageModel = class {
     const add = (setting, name) => {
       if (setting !== void 0)
         unsupported.push({
-          type: "unsupported-setting",
-          setting: name,
+          type: "unsupported",
+          feature: name,
           details: `Codex CLI does not support ${name}; it will be ignored.`
         });
     };
@@ -748,15 +748,21 @@ var CodexCliLanguageModel = class {
   extractUsage(evt) {
     const reported = evt.usage;
     if (!reported) return void 0;
-    const inputTokens = reported.input_tokens ?? 0;
-    const outputTokens = reported.output_tokens ?? 0;
+    const inputTotal = reported.input_tokens ?? 0;
+    const outputTotal = reported.output_tokens ?? 0;
     const cachedInputTokens = reported.cached_input_tokens ?? 0;
     return {
-      inputTokens,
-      outputTokens,
-      // totalTokens should not double-count cached tokens; track cached separately
-      totalTokens: inputTokens + outputTokens,
-      cachedInputTokens
+      inputTokens: {
+        total: inputTotal,
+        noCache: inputTotal - cachedInputTokens,
+        cacheRead: cachedInputTokens,
+        cacheWrite: void 0
+      },
+      outputTokens: {
+        total: outputTotal,
+        text: outputTotal,
+        reasoning: void 0
+      }
     };
   }
   getToolName(item) {
@@ -954,7 +960,10 @@ var CodexCliLanguageModel = class {
       `[codex-cli] Executing Codex CLI: ${cmd} with ${args.length} arguments, cwd: ${cwd ?? "default"}`
     );
     let text = "";
-    const usage = { inputTokens: 0, outputTokens: 0, totalTokens: 0 };
+    let usage = {
+      inputTokens: { total: 0, noCache: 0, cacheRead: 0, cacheWrite: void 0 },
+      outputTokens: { total: 0, text: 0, reasoning: void 0 }
+    };
     const finishReason = "stop";
     const startTime = Date.now();
     const child = spawn(cmd, args, { env, cwd, stdio: ["ignore", "pipe", "pipe"] });
@@ -1000,9 +1009,7 @@ var CodexCliLanguageModel = class {
             if (event.type === "turn.completed") {
               const usageEvent = this.extractUsage(event);
               if (usageEvent) {
-                usage.inputTokens = usageEvent.inputTokens;
-                usage.outputTokens = usageEvent.outputTokens;
-                usage.totalTokens = usageEvent.totalTokens;
+                usage = usageEvent;
               }
             }
             if (event.type === "item.completed" && this.getItemType(event.item) === "assistant_message" && typeof event.item?.text === "string") {
@@ -1037,11 +1044,12 @@ var CodexCliLanguageModel = class {
               );
               return;
             }
+            const totalTokens = (usage.inputTokens.total ?? 0) + (usage.outputTokens.total ?? 0);
             this.logger.info(
-              `[codex-cli] Request completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${usage.totalTokens}`
+              `[codex-cli] Request completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${totalTokens}`
             );
             this.logger.debug(
-              `[codex-cli] Token usage - Input: ${usage.inputTokens}, Output: ${usage.outputTokens}, Total: ${usage.totalTokens}`
+              `[codex-cli] Token usage - Input: ${usage.inputTokens.total ?? 0}, Output: ${usage.outputTokens.total ?? 0}, Total: ${totalTokens}`
             );
             resolve();
           } else {
@@ -1269,12 +1277,16 @@ var CodexCliLanguageModel = class {
             controller.enqueue({ type: "text-delta", id: textId, delta: finalText });
             controller.enqueue({ type: "text-end", id: textId });
           }
-          const usageSummary = lastUsage ?? { inputTokens: 0, outputTokens: 0, totalTokens: 0 };
+          const usageSummary = lastUsage ?? {
+            inputTokens: { total: 0, noCache: 0, cacheRead: 0, cacheWrite: void 0 },
+            outputTokens: { total: 0, text: 0, reasoning: void 0 }
+          };
+          const totalTokens = (usageSummary.inputTokens.total ?? 0) + (usageSummary.outputTokens.total ?? 0);
           this.logger.info(
-            `[codex-cli] Stream completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${usageSummary.totalTokens}`
+            `[codex-cli] Stream completed - Session: ${this.sessionId ?? "N/A"}, Duration: ${duration}ms, Tokens: ${totalTokens}`
           );
           this.logger.debug(
-            `[codex-cli] Token usage - Input: ${usageSummary.inputTokens}, Output: ${usageSummary.outputTokens}, Total: ${usageSummary.totalTokens}`
+            `[codex-cli] Token usage - Input: ${usageSummary.inputTokens.total ?? 0}, Output: ${usageSummary.outputTokens.total ?? 0}, Total: ${totalTokens}`
           );
           controller.enqueue({
             type: "finish",
@@ -1378,8 +1390,8 @@ function createCodexCli(options = {}) {
   };
   provider.languageModel = createModel;
   provider.chat = createModel;
-  provider.textEmbeddingModel = ((modelId) => {
-    throw new NoSuchModelError({ modelId, modelType: "textEmbeddingModel" });
+  provider.embeddingModel = ((modelId) => {
+    throw new NoSuchModelError({ modelId, modelType: "embeddingModel" });
   });
   provider.imageModel = ((modelId) => {
     throw new NoSuchModelError({ modelId, modelType: "imageModel" });

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "ai-sdk-provider-codex-cli",
-  "version": "0.7.0",
-  "description": "AI SDK v5 provider for OpenAI Codex CLI with native JSON Schema support",
+  "version": "1.0.0-beta.1",
+  "description": "AI SDK v6 provider for OpenAI Codex CLI with native JSON Schema support",
   "keywords": [
     "ai-sdk",
     "codex",
@@ -59,8 +59,8 @@
     "test:coverage": "vitest run --coverage"
   },
   "dependencies": {
-    "@ai-sdk/provider": "2.0.0",
-    "@ai-sdk/provider-utils": "3.0.3",
+    "@ai-sdk/provider": "3.0.0-beta.26",
+    "@ai-sdk/provider-utils": "4.0.0-beta.51",
     "jsonc-parser": "^3.3.1"
   },
   "optionalDependencies": {
@@ -70,7 +70,7 @@
     "@eslint/js": "^9.14.0",
     "@types/node": "20.19.6",
     "@vitest/coverage-v8": "^3.2.4",
-    "ai": "5.0.93",
+    "ai": "6.0.0-beta.156",
     "eslint": "^9.14.0",
     "prettier": "^3.3.3",
     "tsup": "8.5.0",