refactor: remove CLI backends, add Chinese README

Remove Claude Code CLI and Codex CLI backends — they depend on unstable CLI flags and macOS Keychain ACLs. API backend is now the only path, default provider is Anthropic. Replace English README with Chinese version focused on API usage. Add .agent/ to .gitignore. Change-Id: I7b2488d9bca23e753cc3d461f5f7bc39052dcbef
2026-07-22 06:53:43 +00:00 · 2026-04-26 09:57:30 +08:00 · 2026-04-26 09:57:30 +08:00 · 9e369ae1fa
commit 9e369ae1fa
parent b7d1a25543
6 changed files with 119 additions and 255 deletions
--- a/.gitignore
+++ b/.gitignore
@ -13,3 +13,6 @@ cache.json
 node_modules/
 dist/
 *.log
+
+# Agent state
+.agent/
--- a/README.md
+++ b/README.md
@ -1,183 +1,125 @@
 # Obsidian Parallel Reader

-An Obsidian plugin that splits any long note into a left/right parallel-reading layout: the original content stays on the left; LLM-generated section summaries appear on the right. Scrolling the source highlights the matching summary card; clicking a summary jumps the source editor.
+一个 Obsidian 插件：左边原文、右边 AI 摘要卡片，滚动联动、点击跳转。

-Inspired by the reading workflow demo in [this Bilibili video](https://www.bilibili.com/video/BV1FxoGBVETm/).
+灵感来自 [这个 B 站视频](https://www.bilibili.com/video/BV1FxoGBVETm/) 的阅读工作流演示。

-## Features
+## 功能

- **Self-adaptive segmentation** — the LLM decides natural topic boundaries, no need to match markdown headings. Short adjacent sections merge; long ones split.
- **Anchor-based line resolution** — summaries carry a verbatim quote from the source; the plugin finds the line number by `indexOf` (with graceful fallbacks).
- **Scroll-sync highlighting** — scrolling the source editor updates the active card in the summary pane.
- **Right-click → copy** — Markdown / plain text / anchor quote / jump to source, as a native Obsidian context menu.
- **Markdown rendering** — summaries render through Obsidian's official `MarkdownRenderer`, so tables, bold, code, wikilinks all work.
- **Persistent bounded cache** — generated summaries are cached by SHA-1 of source content plus generation settings in a separate `cache.json`. Reopening a note shows the cached view instantly; edits or model changes invalidate the cache (shows a stale banner). The cache has an LRU-style entry limit.
- **Panel actions** — regenerate, copy all Markdown, or export the current comparison note directly from the right pane.
- **Vault lifecycle aware** — cached summaries follow Markdown file renames and are removed when the source note is deleted.
- **Configurable input limit** — default 20,000 characters, adjustable for long-context models and included in the cache fingerprint.
- **Prompt controls** — choose summary language, card count range, and an optional custom system prompt. Prompt settings are included in the cache fingerprint.
- **Localized UI** — plugin commands, settings, pane text, and notices support Chinese and English.
- **Multiple model access paths** — Claude Code CLI, Codex CLI, direct provider APIs, and OpenAI/Anthropic-compatible proxies. The API backend follows the same lightweight `provider/model + base URL + protocol` idea as OpenClaw.
+- **自适应切段** — LLM 自行判断主题边界，不依赖 markdown 标题。短段合并、长段拆分。
+- **Anchor 定位** — 每张摘要卡片携带原文逐字引用，插件通过 `indexOf` 定位行号（含多级容错回退）。
+- **滚动联动** — 滚动编辑器时，右侧对应卡片自动高亮。
+- **右键菜单** — 复制 Markdown / 纯文本 / anchor 引用 / 跳转原文 / 编辑 / 删除卡片。
+- **Markdown 渲染** — 摘要通过 Obsidian 的 `MarkdownRenderer` 渲染，表格、加粗、代码、wikilink 均正常显示。
+- **流式输出** — 支持 OpenAI Chat 和 Anthropic API 的 SSE 流式响应，生成时实时显示进度。
+- **持久化缓存** — 生成结果按源文件 SHA-1 + 生成配置缓存在 `cache.json`，重新打开秒出；内容或配置变更后显示"已过期"提示。
+- **面板操作** — 重新生成、复制全部 Markdown、导出到 Vault。
+- **文件感知** — 缓存跟随文件重命名，源文件删除时自动清理。
+- **Prompt 控制** — 可选输出语言、卡片数量范围、自定义 system prompt。
+- **中英双语 UI** — 命令、设置、面板文案全部支持中文和英文。
+- **多 Provider 支持** — Anthropic、OpenAI、Gemini、OpenRouter、Groq、DeepSeek、Moonshot、千帆、MiniMax、xAI、Mistral、Cerebras、智谱、Ollama、LM Studio 及自定义兼容端点。

-## Installation
+## 安装

-### Manual install (until this lands in Community Plugins)
+### 手动安装

-1. Clone this repo into `.obsidian/plugins/parallel-reader/` in your vault, or symlink it there:
-   ```bash
-   ln -s /path/to/obsidian-parallel-reader /path/to/YourVault/.obsidian/plugins/parallel-reader
-   ```
-2. In Obsidian: **Settings → Community plugins → Installed plugins** → enable **Parallel Reader**.
+1. 下载 [最新 Release](https://github.com/fancive/obsidian-parallel-reader/releases) 中的 `main.js`、`manifest.json`、`styles.css`
+2. 在 Vault 的 `.obsidian/plugins/` 目录下创建 `parallel-reader` 文件夹
+3. 将三个文件放入该文件夹
+4. Obsidian → **设置 → 第三方插件 → 已安装插件** → 启用 **Parallel Reader**

-### Pick a backend
+### 配置 Provider

-| Backend | Auth | Cost | Notes |
-|---------|------|------|-------|
-| **Claude Code CLI** | OAuth in macOS Keychain | Included in Claude subscription | ⚠️ Keychain ACL may block Obsidian subprocess reads — may not work from GUI |
-| **Codex CLI** | File-based token (`~/.codex/auth.json`) | Included in ChatGPT Plus | ✅ Works from Obsidian GUI; recommended |
-| **API / Provider** | API key or local server | Provider billing / local | Supports Anthropic, OpenAI, Gemini, OpenRouter, Groq, DeepSeek, Moonshot, Qianfan, MiniMax, xAI, Mistral, Cerebras, Z.AI, Ollama, LM Studio, and custom compatible endpoints |
+在插件设置中选择一个 Provider preset，填入 API Key 和模型 ID 即可。

-### API provider mode
+| Provider | API 格式 | 说明 |
+|----------|----------|------|
+| Anthropic | `anthropic-messages` | 默认 preset，推荐 |
+| OpenAI | `openai-chat` | Chat Completions |
+| Google Gemini | `google-generative-ai` | generateContent |
+| OpenRouter / Groq / DeepSeek / Moonshot 等 | `openai-chat` | OpenAI 兼容格式 |
+| Ollama / LM Studio | `openai-chat` | 本地模型，无需 API Key |
+| 自定义端点 | 任意 | 填写 Base URL 即可 |

-In plugin settings, choose **Backend → API / Provider**, then select a provider preset. Presets fill the protocol, base URL, auth header, and environment variable name; you still control the model ID.
+Model ID 支持 `provider/model` 写法（如 `anthropic/claude-sonnet-4-6`），匹配当前 preset 时自动剥离前缀。

-Supported protocol formats:
+## 命令

-| Format | Use it for |
-|--------|------------|
-| `anthropic-messages` | Anthropic API and Anthropic-compatible proxies |
-| `openai-chat` | OpenAI-compatible `/chat/completions` providers and local proxies |
-| `openai-responses` | OpenAI `/responses` endpoint |
-| `google-generative-ai` | Gemini `generateContent` API |
+| 命令 | 说明 |
+|------|------|
+| 为当前笔记生成对照笔记（缓存优先） | 有缓存直接显示，否则调用 LLM |
+| 强制重新生成（绕过缓存） | 忽略缓存重新调用 LLM |
+| 打开对照笔记面板 | 打开右侧面板 |
+| 导出当前对照笔记到 Vault | 保存为 Markdown 文件 |
+| 复制当前对照笔记 Markdown | 复制到剪贴板 |
+| 取消当前对照笔记生成 | 取消正在进行的生成 |
+| `Alt+↑` / `Alt+↓` | 在摘要卡片间切换 |
+| `Enter`（在摘要面板内） | 跳转到当前卡片对应的原文位置 |

-Model IDs accept OpenClaw-style `provider/model` values. If the prefix matches the selected preset, the plugin strips it before the request. For example:
+## 交互

- Preset `anthropic`, model `anthropic/claude-sonnet-4-6` → sends `claude-sonnet-4-6`
- Preset `openrouter`, model `openrouter/anthropic/claude-sonnet-4-5` → sends `anthropic/claude-sonnet-4-5`
- Preset `ollama`, model `llama3.3` → sends `llama3.3`
+| 操作 | 效果 |
+|------|------|
+| 点击卡片 | 跳转到原文对应位置 |
+| 右键卡片 | 上下文菜单 |
+| 滚动编辑器 | 右侧卡片自动高亮 |
+| 拖选文字 | 正常选择文本（不触发跳转） |
+| 文件右键菜单 | 生成 / 重新生成 / 清除缓存 |
+| Ribbon 图标 | 打开对照面板 |

-For Cloudflare AI Gateway or other proxies, use **额外 headers** with either JSON:
-
-```json
-{"cf-aig-authorization":"Bearer ..."}
-```
-
-or one header per line:
-
-```text
-cf-aig-authorization: Bearer ...
-```
-
-### Configure the CLI path
-
-Obsidian launched from Finder does **not** inherit your shell `PATH`. Create a stable symlink in `~/bin/` and point the plugin setting at it:
-
-```bash
-mkdir -p ~/bin
-# Claude Code (Mach-O binary, no Node needed)
-ln -sf "$(readlink -f "$(which claude)")" ~/bin/claude
-# Codex (Node script — use the wrapper pattern below instead)
-cat > ~/bin/codex <<'EOF'
-#!/bin/bash
-NODE="/path/to/your/node"
-CODEX_JS="/path/to/@openai/codex/bin/codex.js"
-exec "$NODE" "$CODEX_JS" "$@"
-EOF
-chmod +x ~/bin/codex
-```
-
-Then in the plugin settings, paste `/Users/you/bin/codex` (or `~/bin/claude`) into **CLI 路径**, and click **Test** to verify.
-
-## Commands
-
-| Command | What it does |
-|---------|--------------|
-| `为当前笔记生成对照笔记（缓存优先）` | Show cached summary if present, otherwise call LLM |
-| `强制重新生成（绕过缓存）` | Re-run the LLM even if cache matches |
-| `打开对照笔记面板` | Open right-pane view (loads cache if present) |
-| `导出当前对照笔记到 Vault` | Save the current comparison note under the configured export folder |
-| `复制当前对照笔记 Markdown` | Copy the current comparison note as Markdown |
-| `取消当前对照笔记生成` | Mark the active generation job as cancelled |
-| `清除当前笔记的缓存` | Drop the cache entry for the active note |
-| `清除所有缓存` | Wipe all cached summaries |
-| `聚焦上一张摘要卡片` / `Focus previous summary card` | Move the active summary card upward |
-| `聚焦下一张摘要卡片` / `Focus next summary card` | Move the active summary card downward |
-| `跳转到当前摘要卡片原文` / `Jump current summary card to source` | Jump the source editor to the active card |
-
-## Interaction
-
-| Action | Effect |
-|--------|--------|
-| Left-click a card body | Scroll source editor to that section |
-| Right-click a card | Context menu: Copy Markdown / Copy plain text / Copy anchor / Jump / Edit / Delete card |
-| Header icon buttons | Regenerate or cancel / copy all Markdown / export to Vault |
-| File context menu | Generate / force regenerate / clear cache for a Markdown file |
-| Ribbon icon | Open the comparison pane for the active note |
-| `Alt+↑` / `Alt+↓` | Move between summary cards |
-| `Enter` in the summary pane | Jump to the active card's source line |
-| Drag to select text | Normal text selection (does not trigger jump) |
-| Scroll source editor | Active card gets highlighted on the right |
-
-## Development
+## 开发

 ```bash
 npm install
-npm run dev       # watch main.ts + src/**/*.ts and rebuild main.js
-npm run build     # production bundle for Obsidian
-npm run typecheck
-npm test
-node --check main.js
+npm run dev       # 监听 main.ts + src/**/*.ts，自动重新构建
+npm run build     # 生产构建
+npm run typecheck # TypeScript 类型检查（strict 模式）
+npm run lint      # Biome lint
+npm test          # 构建 + 类型检查 + 测试
 ```

-### Code layout
+### 项目结构

-The plugin keeps `main.js` as the generated Obsidian runtime bundle. Edit `main.ts` and `src/**/*.ts`; esbuild bundles them back into root `main.js`.
+| 文件 | 职责 |
+|------|------|
+| `main.ts` | 插件生命周期、命令注册、缓存管理、滚动联动 |
+| `src/view.ts` | 右侧面板视图、卡片渲染、键盘导航、导出 |
+| `src/modal.ts` | 卡片编辑弹窗 |
+| `src/settings-tab.ts` | 设置面板 |
+| `src/providers.ts` | API 请求/响应适配器（Anthropic、OpenAI、Gemini） |
+| `src/streaming.ts` | SSE 流式解析 |
+| `src/prompt.ts` | Prompt 构建、语言控制、自定义 prompt 模板 |
+| `src/schema.ts` | JSON 提取、卡片数据归一化、结构化输出 schema |
+| `src/anchor.ts` | Anchor 到行号的匹配（含容错） |
+| `src/settings.ts` | 默认值、Provider preset、缓存指纹 |
+| `src/i18n.ts` | 中英文 UI 翻译 |
+| `src/cache.ts` / `src/cards.ts` / `src/navigation.ts` / `src/scroll.ts` / `src/vault.ts` / `src/markdown.ts` | 各种纯函数工具模块 |

-| File | Responsibility |
-|------|----------------|
-| `main.ts` | Obsidian lifecycle, commands, right-pane view, settings tab orchestration |
-| `src/anchor.ts` | Anchor-to-line matching and whitespace-normalized fallback |
-| `src/cache.ts` | Cache entry touch semantics and compact cache-file serialization |
-| `src/cards.ts` | Card list edit/delete helpers |
-| `src/i18n.ts` | Chinese/English UI strings and translation helper |
-| `src/navigation.ts` | Summary-card keyboard navigation helpers |
-| `src/prompt.ts` | Prompt construction, language controls, and custom prompt templating |
-| `src/settings.ts` | Defaults, provider presets, settings normalization, cache fingerprinting and pruning |
-| `src/vault.ts` | Vault path normalization and recursive folder creation |
-| `src/schema.ts` | JSON extraction, card payload normalization, structured-output schemas |
-| `src/scroll.ts` | Scroll-sync requestAnimationFrame throttling helper |
-| `src/providers.ts` | API provider request/response adapters |
-| `src/generation-job-manager.ts` | Per-file generation state, cancellation, and error classification |
-| `src/markdown.ts` | Card-to-Markdown/plain-text serialization |
+## 设计

-## Design notes
-
-The LLM is asked to return JSON with this shape:
+LLM 返回如下 JSON：

 ```json
 {
  "cards": [
    {
-      "title": "3-10 character short heading",
-      "anchor": "40-80 characters, verbatim from source, used for line resolution",
-      "gist": "20-40 character one-line summary serving as a lead-in",
-      "bullets": ["3-6 supporting details, 20-50 chars each"]
+      "title": "3-10 字短标题",
+      "anchor": "40-80 字，从原文逐字复制，用于定位",
+      "gist": "20-40 字一句话领读",
+      "bullets": ["3-6 条支撑要点，每条 20-50 字"]
    }
  ]
 }
 ```

- The **anchor** is the key mechanism that keeps scroll-sync working without relying on markdown headings. It's copied verbatim so `content.indexOf(anchor)` finds the origin line; graceful fallbacks (prefix trimming, whitespace normalization) handle minor LLM drift.
- The **gist + bullets** dual structure was chosen after several iterations — pure prose felt wall-of-text, pure bullets felt fragmented. A one-line lead-in before bullets gives both overview and scannable detail.
- Output is rendered via `MarkdownRenderer.render()` so any tables / bold / code / wikilinks in the LLM response render natively.
+- **anchor** 是滚动联动的核心机制，通过 `content.indexOf(anchor)` 定位行号，不依赖 markdown 标题。
+- **gist + bullets** 双层结构在多次迭代后确定 — 纯散文太密、纯列表太碎，一句话导读 + bullet 兼顾概览和细节。

-## Known limitations
+## 已知限制

- **Document size cap**: defaults to 20,000 characters. Longer notes are truncated unless you raise **最大输入字符数** in settings.
- **Claude Code CLI auth**: Obsidian subprocess reads of the macOS Keychain `Claude Code-credentials` service may fail due to code-signing ACL. Codex backend doesn't have this issue.
- **No streaming**: generation is a single batch call; expect ~5-15s wait depending on document length and backend.
- **Provider feature gaps**: some OpenAI-compatible providers reject optional fields or use provider-specific model IDs. If a request fails, verify the model ID and switch between `openai-chat`, `openai-responses`, or the provider's native format.
- **Preview mode**: scroll-sync uses Obsidian's `setEphemeralState` which works in reading mode too, but accuracy depends on how CodeMirror resolves line positions.
+- **文档长度**：默认 20,000 字符，超长笔记会截断（可在设置中调大）。
+- **Provider 差异**：部分 OpenAI 兼容 provider 可能不支持某些可选字段，请根据报错切换 API 格式。
+- **Preview 模式**：滚动联动在阅读模式下也能工作，但精度取决于 CodeMirror 的行位置计算。

 ## License

--- a/main.js
+++ b/main.js
--- a/main.ts
+++ b/main.ts
@ -3,7 +3,7 @@ import { MarkdownView, Notice, Plugin, requestUrl, TFile } from 'obsidian';
 import { findLineForAnchor } from './src/anchor';
 import { serializeCacheFile, shouldConfirmRegenerate, touchCacheEntry } from './src/cache';
 import { activeIndexAfterCardDelete, removeCardAt, updateCardAt } from './src/cards';
-import { resolveCliPath, summarizeViaClaudeCode, summarizeViaCodex } from './src/cli';
+import { resolveCliPath } from './src/cli';
 import {
  classifyGenerationError,
  type GenerationJob,
@ -57,26 +57,15 @@ async function summarizeDocument(
 ) {
  const { system, user } = buildPrompts(content, settings);
  let cards: RawCard[];
-  const useStreaming = isApiBackend(settings.backend) && supportsStreaming(settings);
+  const useStreaming = supportsStreaming(settings);
  const abortController = useStreaming ? new AbortController() : null;
  if (abortController) {
    job.onCancel(() => abortController.abort());
  }
-  switch (settings.backend) {
-    case 'codex':
-      cards = await summarizeViaCodex(system, user, settings, job);
-      break;
-    case 'api':
-    case 'anthropic-api':
-      if (useStreaming) {
-        cards = await summarizeViaApiStreaming(system, user, settings, onStreamProgress, abortController!.signal);
-      } else {
-        cards = await summarizeViaApi(requestUrl, system, user, settings);
-      }
-      break;
-    default:
-      cards = await summarizeViaClaudeCode(system, user, settings, job);
-      break;
+  if (useStreaming) {
+    cards = await summarizeViaApiStreaming(system, user, settings, onStreamProgress, abortController!.signal);
+  } else {
+    cards = await summarizeViaApi(requestUrl, system, user, settings);
  }
  const resolved: ResolvedCard[] = cards.map((c) => ({
    title: c.title,
--- a/src/settings-tab.ts
+++ b/src/settings-tab.ts
@ -1,7 +1,6 @@
 'use strict';

 import { type App, Notice, type Plugin, PluginSettingTab, requestUrl, Setting } from 'obsidian';
-import { resolveCliPath, runCli } from './cli';
 import { testApiBackend } from './providers';
 import {
  API_AUTH_TYPES,
@ -12,27 +11,13 @@ import {
  DEFAULT_SETTINGS,
  getApiFormat,
  getApiPreset,
-  isApiBackend,
  PROMPT_LANGUAGES,
  UI_LANGUAGES,
 } from './settings';
 import type { PluginHost, PluginSettings } from './types';

 async function testBackend(settings: PluginSettings) {
-  if (settings.backend === 'codex') {
-    const cmd = resolveCliPath('codex', settings.cliPath);
-    const { stdout } = await runCli(cmd, ['--version'], '', 10000);
-    return `codex @ ${cmd}\n${stdout.trim()}`;
-  }
-  if (settings.backend === 'claude-code') {
-    const cmd = resolveCliPath('claude', settings.cliPath);
-    const { stdout } = await runCli(cmd, ['--version'], '', 10000);
-    return `claude @ ${cmd}\n${stdout.trim()}`;
-  }
-  if (isApiBackend(settings.backend)) {
-    return testApiBackend(requestUrl, settings);
-  }
-  throw new Error('Unknown backend: ' + settings.backend);
+  return testApiBackend(requestUrl, settings);
 }

 export class ParallelReaderSettingTab extends PluginSettingTab {
@ -63,54 +48,9 @@ export class ParallelReaderSettingTab extends PluginSettingTab {
        });
      });

-    new Setting(containerEl)
-      .setName(tr('settingBackendName'))
-      .setDesc(tr('settingBackendDesc'))
-      .addDropdown((d) =>
-        d
-          .addOption('claude-code', 'Claude Code CLI')
-          .addOption('codex', 'Codex CLI')
-          .addOption('api', 'API / Provider')
-          .addOption('anthropic-api', 'Anthropic API (legacy)')
-          .setValue(this.plugin.settings.backend)
-          .onChange(async (v) => {
-            this.plugin.settings.backend = v;
-            if (v === 'api' && !this.plugin.settings.apiBaseUrl) {
-              applyApiProviderPreset(this.plugin.settings, this.plugin.settings.apiProvider || 'anthropic');
-            }
-            await this.plugin.saveSettings();
-            this.display();
-          }),
-      );
-
-    const apiBackend = isApiBackend(this.plugin.settings.backend);
-
-    if (!apiBackend) {
-      new Setting(containerEl)
-        .setName(tr('settingCliPathName'))
-        .setDesc(tr('settingCliPathDesc'))
-        .addText((t) =>
-          t
-            .setPlaceholder(tr('settingCliPathPlaceholder'))
-            .setValue(this.plugin.settings.cliPath)
-            .onChange(async (v) => {
-              this.plugin.settings.cliPath = v.trim();
-              this.plugin.saveSettingsDebounced();
-            }),
-        );
-
-      new Setting(containerEl).setName(tr('settingCliTimeoutName')).addText((t) =>
-        t.setValue(String(this.plugin.settings.cliTimeoutMs)).onChange(async (v) => {
-          const n = parseInt(v, 10);
-          if (!Number.isNaN(n) && n > 0) {
-            this.plugin.settings.cliTimeoutMs = n;
-            this.plugin.saveSettingsDebounced();
-          }
-        }),
-      );
-    } else {
-      containerEl.createEl('h3', { text: tr('apiProviderHeader') });
+    containerEl.createEl('h3', { text: tr('apiProviderHeader') });

+    {
      const preset = getApiPreset(this.plugin.settings);
      new Setting(containerEl)
        .setName(tr('settingProviderPresetName'))
@ -233,13 +173,13 @@ export class ParallelReaderSettingTab extends PluginSettingTab {

    new Setting(containerEl)
      .setName(tr('settingModelName'))
-      .setDesc(apiBackend ? tr('settingModelDescApi') : tr('settingModelDescCli'))
+      .setDesc(tr('settingModelDescApi'))
      .addText((t) =>
        t
-          .setPlaceholder(apiBackend ? getApiPreset(this.plugin.settings).model || 'model-id' : DEFAULT_SETTINGS.model)
+          .setPlaceholder(getApiPreset(this.plugin.settings).model || 'model-id')
          .setValue(this.plugin.settings.model)
          .onChange(async (v) => {
-            this.plugin.settings.model = v.trim() || (apiBackend ? '' : DEFAULT_SETTINGS.model);
+            this.plugin.settings.model = v.trim();
            this.plugin.saveSettingsDebounced();
          }),
      );
@ -319,7 +259,7 @@ export class ParallelReaderSettingTab extends PluginSettingTab {

    new Setting(containerEl)
      .setName(tr('settingTestBackendName'))
-      .setDesc(apiBackend ? tr('settingTestBackendDescApi') : tr('settingTestBackendDescCli'))
+      .setDesc(tr('settingTestBackendDescApi'))
      .addButton((b) =>
        b.setButtonText('Test').onClick(async () => {
          try {
--- a/src/settings.ts
+++ b/src/settings.ts
@ -21,7 +21,7 @@ export const UI_LANGUAGES = {

 export const DEFAULT_SETTINGS: PluginSettings = {
  uiLanguage: 'auto',
-  backend: 'claude-code',
+  backend: 'api',
  cliPath: '',
  apiProvider: 'anthropic',
  apiFormat: 'anthropic-messages',