felo-ai 0.2.0 → 0.2.2

package/docs/FAQ.md

@@ -191,10 +191,10 @@ Use Felo to find [query]
 
 **A:** Yes! Fully supports multi-language queries.
 
-Ask in any language, get answers in that language. Even mixed-language queries work:
-- "React 在中国的使用情况"
-- "東京で中国語が通じる病院"
-- "Python vs Java 哪个更适合初学者？"
+Ask in any language, get answers in that language. Even mixed-language queries work, for example:
+- "React adoption in China"
+- "Hospitals in Tokyo with Chinese-speaking staff"
+- "Python vs Java for beginners?"
 
 ### Q: What's the response format?
 

package/felo-search/README.md

@@ -112,8 +112,8 @@ You: What's the weather in Tokyo today?
 Claude: Currently 15°C in Tokyo with partly cloudy skies. High of 18°C,
         low of 12°C. 20% chance of rain.
 
-You: 杭州明天天气怎么样？
-Claude: 杭州明天多云转晴，气温 8-16°C，东北风 3-4 级。
+You: What's the weather in Hangzhou tomorrow?
+Claude: Hangzhou tomorrow: partly cloudy to clear, 8-16°C, NE wind 3-4.
 ```
 
 **Restaurants & Food**
@@ -124,11 +124,11 @@ Claude: Top 3 ramen shops in Osaka:
         2. Kamukura - Rich chicken broth, Michelin Bib Gourmand
         3. Menya Joroku - Tsukemen specialist
 
-You: 上海哪里有好吃的小笼包？
-Claude: 上海小笼包推荐：
-        1. 南翔馒头店 - 豫园老字号，皮薄汁多
-        2. 佳家汤包 - 黄河路网红店，性价比高
-        3. 鼎泰丰 - 台湾品牌，品质稳定
+You: Where to find the best xiaolongbao in Shanghai?
+Claude: Shanghai xiaolongbao recommendations:
+        1. Nanxiang Steamed Bun - Old brand near Yuyuan, thin skin, juicy
+        2. Jiajia Soup Dumplings - Popular on Huanghe Road, good value
+        3. Din Tai Fung - Taiwan brand, consistent quality
 ```
 
 **Shopping & Prices**
@@ -139,11 +139,11 @@ Claude: iPhone 15 Pro (128GB) prices:
         - Amazon: $979 (save $20)
         - Best Buy: $999 (with trade-in deals)
 
-You: MacBook Air M3 多少钱？
-Claude: MacBook Air M3 (13 英寸) 价格：
-        - 官网: ¥9,499 (256GB)
-        - 京东: ¥9,299 (有优惠券)
-        - 天猫: ¥9,399 (赠品多)
+You: MacBook Air M3 price
+Claude: MacBook Air M3 (13") prices:
+        - Apple: $1,099 (256GB)
+        - Amazon: $1,049 (with coupons)
+        - Best Buy: $1,079 (with perks)
 ```
 
 **Travel**
@@ -155,12 +155,12 @@ Claude: Kyoto weekend highlights:
         - Nishiki Market food tour (closes 6pm)
         - Gion district evening walk (geisha spotting)
 
-You: 台北有什么好玩的地方？
-Claude: 台北必去景点：
-        - 台北 101 - 观景台看夜景
-        - 故宫博物院 - 中华文物精品
-        - 士林夜市 - 美食天堂
-        - 九份老街 - 怀旧山城
+You: What to do in Taipei?
+Claude: Taipei must-sees:
+        - Taipei 101 - Observatory for night views
+        - National Palace Museum - Chinese art and artifacts
+        - Shilin Night Market - Food and snacks
+        - Jiufen Old Street - Nostalgic mountain town
 ```
 
 ### Developer Scenarios
@@ -174,12 +174,12 @@ Claude: React 19 key features (released Dec 2024):
         - Actions for form handling
         - use() hook for async data
 
-You: Next.js 15 有什么新功能？
-Claude: Next.js 15 主要更新：
-        - Turbopack 稳定版（构建速度提升 5x）
-        - 部分预渲染（PPR）
-        - 改进的缓存控制
-        - React 19 支持
+You: Next.js 15 new features
+Claude: Next.js 15 main updates:
+        - Turbopack stable (5x faster builds)
+        - Partial Prerendering (PPR)
+        - Improved cache control
+        - React 19 support
 ```
 
 **Library Comparison**
@@ -194,15 +194,15 @@ Claude: Vite vs Webpack (2024):
         Use Vite for: New projects, fast iteration
         Use Webpack for: Complex builds, legacy support
 
-You: Vue 3 和 React 哪个更适合新项目？
-Claude: Vue 3 vs React 对比：
+You: Vue 3 vs React for new projects
+Claude: Vue 3 vs React:
 
-        学习曲线: Vue 更平缓，React 需要 JSX
-        性能: 两者相当，Vue 略优
-        生态: React 更丰富，Vue 够用
-        就业: React 岗位更多
+        Learning curve: Vue gentler, React needs JSX
+        Performance: Similar, Vue slightly ahead
+        Ecosystem: React larger, Vue sufficient
+        Jobs: More React positions
 
-        建议: 团队熟悉 React 选 React，新手选 Vue
+        Suggestion: React if team knows it, Vue for newcomers
 ```
 
 **Tech Trends**
@@ -214,109 +214,54 @@ Claude: AI developments (Jan 2026):
         - Anthropic Claude 4 with extended context
         - Meta's open-source LLaMA 4
 
-You: 最近有什么重要的技术新闻？
-Claude: 近期科技动态（2026年2月）：
-        - 苹果发布 Vision Pro 2 代
-        - 英伟达 B100 GPU 开始量产
-        - 特斯拉 FSD v13 正式推送
-        - 微软 Windows 12 预览版发布
+You: Important tech news recently
+Claude: Recent tech (Feb 2026):
+        - Apple Vision Pro 2 announced
+        - NVIDIA B100 GPU in production
+        - Tesla FSD v13 rollout
+        - Microsoft Windows 12 preview
 ```
 
 ### Multi-language Queries
 
-**Chinese (Simplified)**
-```
-You: 人工智能最新进展
-Claude: AI 最新进展（2026年）：
-        - 大模型参数突破 10 万亿
-        - 多模态理解接近人类水平
-        - AI 编程助手普及率超 60%
-        - 具身智能机器人商业化
-        [来源: 机器之心, 量子位]
-
-You: 深圳有什么好玩的地方
-Claude: 深圳旅游推荐：
-        - 世界之窗 - 微缩景观
-        - 欢乐谷 - 主题乐园
-        - 大梅沙海滨公园 - 免费沙滩
-        - 华侨城创意园 - 文艺打卡
-        [来源: 深圳文旅局, 马蜂窝]
-```
-
-**Chinese (Traditional - Taiwan)**
-```
-You: 台北最近有什麼新鮮事？
-Claude: 台北近期新聞：
-        - 捷運環狀線北環段動工
-        - 大巨蛋正式啟用
-        - 台北燈節 2 月登場
-        - 米其林指南台北版更新
-        [來源: 聯合新聞網, 自由時報]
-
-You: 高雄美食推薦
-Claude: 高雄必吃美食：
-        - 六合夜市 - 木瓜牛奶、海產粥
-        - 瑞豐夜市 - 炸雞排、臭豆腐
-        - 旗津海產街 - 新鮮海鮮
-        - 鹽埕區 - 老字號小吃
-        [來源: 愛食記, 痞客邦]
-```
-
-**Japanese**
-```
-You: 東京で今人気のレストランは？
-Claude: 東京の人気レストラン（2026年）：
-        - Narisawa - モダン日本料理、ミシュラン2つ星
-        - Den - 革新的和食、世界のベストレストラン
-        - Florilège - フレンチ、予約困難
-        - Sushi Saito - 江戸前寿司の最高峰
-        [出典: 食べログ, ミシュランガイド]
-
-You: 最新のAI技術について教えて
-Claude: 最新AI技術（2026年）：
-        - GPT-5の登場で自然言語処理が進化
-        - マルチモーダルAIが主流に
-        - エッジAIの普及が加速
-        - AI規制法が各国で施行
-        [出典: ITmedia, 日経新聞]
-```
-
-**Korean**
-```
-You: 서울 맛집 추천
-Claude: 서울 맛집 추천：
-        - 광장시장 - 전통 한식, 마약김밥
-        - 명동 - 길거리 음식, 쇼핑
-        - 강남 - 고급 레스토랑, 카페
-        - 이태원 - 세계 각국 요리
-        [출처: 망고플레이트, 네이버 블로그]
-
-You: 최신 프로그래밍 트렌드
-Claude: 최신 프로그래밍 트렌드 (2026):
-        - AI 코딩 어시스턴트 보편화
-        - Rust 인기 급상승
-        - 서버리스 아키텍처 확산
-        - WebAssembly 생태계 성장
-        [출처: Stack Overflow, GitHub Trends]
-```
-
-**Mixed Language**
-```
-You: React 在中国的使用情况
-Claude: React 在中国的使用情况：
-        - 市场份额: 约 45%（仅次于 Vue）
-        - 主要用户: 大厂（阿里、字节、腾讯）
-        - 生态: 完善，Ant Design、Taro 等
-        - 趋势: 持续增长，Next.js 受欢迎
-        [来源: State of JS China, 掘金]
-
-You: 日本のAI企業について
-Claude: 日本の主要AI企業：
-        - Preferred Networks - 深層学習フレームワーク
-        - ABEJA - エッジAIプラットフォーム
-        - PKSHA Technology - 自然言語処理
-        - Nextremer - 対話AI
-        [出典: 日経ビジネス, TechCrunch Japan]
+Felo supports Chinese, Japanese, Korean, and English. Ask in any language and get answers in that language. Example in English:
+
+**Latest AI & places**
+```
+You: Latest developments in AI
+Claude: AI developments (2026):
+        - LLM parameters exceed 10 trillion
+        - Multimodal understanding near human level
+        - AI coding assistants used by 60%+ developers
+        - Embodied AI robots commercialized
+        [Sources: industry reports]
+
+You: What to do in Shenzhen
+Claude: Shenzhen recommendations:
+        - Window of the World - Miniature landmarks
+        - Happy Valley - Theme park
+        - Dameisha Beach - Free beach
+        - OCT-LOFT - Arts and culture
+        [Sources: travel guides]
+```
+
+**React adoption & comparisons**
+```
+You: React adoption in China
+Claude: React in China:
+        - Market share: ~45% (second to Vue)
+        - Major users: Alibaba, ByteDance, Tencent
+        - Ecosystem: Ant Design, Taro, etc.
+        - Trend: Growing, Next.js popular
+        [Sources: State of JS China]
+
+You: Python vs Java for beginners
+Claude: Python vs Java:
+        - Learning: Python easier to start
+        - Jobs: Both strong, Java more enterprise
+        - Use Python for: Data, scripting, prototyping
+        - Use Java for: Enterprise, Android, backend
+        [Sources: Stack Overflow, job boards]
 ```
 
 ### Complex Queries
@@ -333,15 +278,15 @@ Claude: AI's impact on software development (2024-2026):
 
         Future: AI pair programming becomes standard practice
 
-You: 量子计算对密码学的影响
-Claude: 量子计算对密码学的影响：
+You: Impact of quantum computing on cryptography
+Claude: Quantum computing and cryptography:
 
-        威胁: RSA、ECC 等传统加密算法将被破解
-        时间线: 预计 2030-2035 年实现
-        应对: 后量子密码学（PQC）标准化
-        进展: NIST 已发布 PQC 算法标准
+        Threat: RSA, ECC and similar algorithms could be broken
+        Timeline: Estimated 2030-2035
+        Response: Post-quantum cryptography (PQC) standardization
+        Progress: NIST has published PQC algorithm standards
 
-        建议: 企业应提前规划密码学迁移
+        Advice: Plan crypto migration in advance
 ```
 
 **Comparisons**
@@ -359,18 +304,18 @@ Claude: AWS vs Azure vs GCP (2026):
         Choose Azure: Microsoft integration, hybrid cloud
         Choose GCP: Data/ML workloads, cost-sensitive
 
-You: 编程语言性能对比：Python vs Go vs Rust
-Claude: Python vs Go vs Rust 性能对比：
+You: Programming language performance: Python vs Go vs Rust
+Claude: Python vs Go vs Rust:
 
-        执行速度: Rust > Go > Python (100x 差距)
-        内存占用: Rust 最小，Python 最大
-        开发效率: Python > Go > Rust
-        学习曲线: Python 最平缓，Rust 最陡
+        Speed: Rust > Go > Python (up to 100x)
+        Memory: Rust lowest, Python highest
+        Productivity: Python > Go > Rust
+        Learning curve: Python gentlest, Rust steepest
 
-        适用场景:
-        - Python: 数据科学、脚本、快速原型
-        - Go: 微服务、云原生、并发
-        - Rust: 系统编程、性能关键、安全
+        Use cases:
+        - Python: Data science, scripting, rapid prototyping
+        - Go: Microservices, cloud-native, concurrency
+        - Rust: Systems programming, performance-critical, safety
 ```
 
 **[See 40+ more examples →](../docs/EXAMPLES.md)**

package/felo-search/SKILL.md

@@ -118,16 +118,16 @@ The API returns JSON with this structure:
 Parse the JSON response and present it to the user in this format:
 
 ```
-## 回答 / Answer
+## Answer
 [Display the answer field]
 
-## 搜索分析 / Query Analysis
-优化后的搜索词: [list query_analysis items]
+## Query Analysis
+Optimized search terms: [list query_analysis items]
 ```
 
 ## Complete Examples
 
-### Example 1: English Query
+### Example 1: Weather query
 
 **User asks:** "What's the weather in Tokyo today?"
 
@@ -156,24 +156,23 @@ curl -s -X POST https://openapi.felo.ai/v2/chat \
 rm -f /tmp/felo_query.json
 ```
 
-### Example 2: Simplified Chinese (简体中文)
+### Example 2: Local news / events
 
-**User asks:** "杭州最近有什么新鲜事？"
+**User asks:** "What's new in Hangzhou recently?"
 
 **Expected response format:**
 ```
-## 回答
-杭州最近的新鲜事包括：亚运会场馆改造完成、西湖景区推出夜游项目、
-新的地铁线路开通等。详细信息...
+## Answer
+Recent news in Hangzhou: Asian Games venue upgrades completed, West Lake night tours launched, new metro lines opened. Details...
 
-## 搜索分析
-优化后的搜索词: 杭州最近新闻, 杭州近期动态, Hangzhou recent news
+## Query Analysis
+Optimized search terms: Hangzhou recent news, Hangzhou events, 杭州 最近 新闻
 ```
 
 **Bash command:**
 ```bash
 cat > /tmp/felo_query.json << 'EOF'
-{"query": "杭州最近有什么新鲜事"}
+{"query": "What's new in Hangzhou recently"}
 EOF
 
 curl -s -X POST https://openapi.felo.ai/v2/chat \
@@ -184,14 +183,14 @@ curl -s -X POST https://openapi.felo.ai/v2/chat \
 rm -f /tmp/felo_query.json
 ```
 
-### Example 3: Traditional Chinese - Taiwan (繁體中文-台灣)
+### Example 3: Travel / things to do
 
-**User asks:** "台北最近有什麼好玩的地方？"
+**User asks:** "What are the best things to do in Taipei?"
 
 **Bash command:**
 ```bash
 cat > /tmp/felo_query.json << 'EOF'
-{"query": "台北最近有什麼好玩的地方"}
+{"query": "What are the best things to do in Taipei"}
 EOF
 
 curl -s -X POST https://openapi.felo.ai/v2/chat \
@@ -202,14 +201,14 @@ curl -s -X POST https://openapi.felo.ai/v2/chat \
 rm -f /tmp/felo_query.json
 ```
 
-### Example 4: Japanese (日本語)
+### Example 4: Restaurants / recommendations
 
-**User asks:** "東京で今人気のレストランは？"
+**User asks:** "Popular restaurants in Tokyo?"
 
 **Bash command:**
 ```bash
 cat > /tmp/felo_query.json << 'EOF'
-{"query": "東京で今人気のレストランは"}
+{"query": "Popular restaurants in Tokyo"}
 EOF
 
 curl -s -X POST https://openapi.felo.ai/v2/chat \

package/felo-slides/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Felo Slides Skill Contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/felo-slides/README.md

@@ -0,0 +1,87 @@
+# Felo Slides Skill for Claude Code
+
+Generate presentation slides with the Felo PPT Task API (asynchronous workflow).
+
+## Features
+
+- Generate a PPT deck from a single prompt
+- Poll task status automatically until completion/failure/timeout
+- Return `ppt_url` immediately when the task is completed (fallback to `live_doc_url`)
+- Return `task_id` for follow-up tracking
+
+## Quick Start
+
+### 1) Install the skill
+
+```bash
+npx @claude/skills add felo-slides
+```
+
+Or install manually from this repository:
+
+```bash
+# Linux/macOS
+cp -r felo-slides ~/.claude/skills/
+
+# Windows (PowerShell)
+Copy-Item -Recurse felo-slides "$env:USERPROFILE\.claude\skills\"
+```
+
+### 2) Configure API key
+
+Create an API key at [felo.ai](https://felo.ai) -> Settings -> API Keys, then set:
+
+```bash
+# Linux/macOS
+export FELO_API_KEY="your-api-key-here"
+```
+
+```powershell
+# Windows PowerShell
+$env:FELO_API_KEY="your-api-key-here"
+```
+
+### 3) Trigger the skill
+
+- Intent trigger: "Create a 10-slide product pitch deck"
+- Explicit trigger: `/felo-slides your topic`
+
+## API Workflow
+
+Based on Felo v2 PPT Task API:
+
+1. Create task: `POST /v2/ppts`
+2. Query status (optional): `GET /v2/tasks/{task_id}/status`
+3. Query historical/result: `GET /v2/tasks/{task_id}/historical`
+
+The skill polls every 10 seconds (max wait 1800 seconds). It stops immediately on `COMPLETED`/`SUCCESS` and returns `ppt_url` (fallback `live_doc_url`).
+
+Internal script example:
+
+```bash
+node felo-slides/scripts/run_ppt_task.mjs --query "Felo product intro, 3 slides" --interval 10 --max-wait 1800
+```
+
+## Troubleshooting
+
+### `FELO_API_KEY` is missing
+
+Set the environment variable and restart the Claude Code session.
+
+### `INVALID_API_KEY`
+
+The key is invalid or revoked. Generate a new key from [felo.ai](https://felo.ai).
+
+### Task keeps running for too long
+
+The task may still be processing. Retry later with the same context, or run the script with `--verbose`.
+
+### Task completed but no `ppt_url` / `live_doc_url`
+
+Use the returned `task_id` to query historical endpoint again.
+
+## Links
+
+- [PPT Task API](https://openapi.felo.ai/docs/api-reference/v2/ppt-tasks.html)
+- [Felo Open Platform](https://openapi.felo.ai/docs/)
+- [Get API Key](https://felo.ai)

package/felo-slides/SKILL.md

@@ -0,0 +1,154 @@
+---
+name: felo-slides
+description: "Generate PPT/slides with Felo PPT Task API in Claude Code. Use when users ask to create/make/generate/export presentations or slide decks, or when explicit commands like /felo-slides are used. Handles API key check, task creation, polling, and final ppt_url output."
+---
+
+# Felo Slides Skill
+
+## When to Use
+
+Trigger this skill for requests about creating presentation files:
+
+- Create/generate slides from a topic or outline
+- Turn notes into a PPT deck
+- Build a presentation with a page count requirement
+- Export presentation content into a shareable slide link
+
+Trigger keywords:
+- Chinese prompts about making slides or presentations
+- English: slides, PPT, presentation deck, generate presentation
+- Explicit commands: `/felo-slides`, "use felo slides"
+
+Do NOT use this skill for:
+- Real-time information lookup (use `felo-search`)
+- Questions about local codebase files
+- Pure text tasks that do not require slide generation
+
+## Setup
+
+### 1. Get API key
+
+1. Visit [felo.ai](https://felo.ai)
+2. Open Settings -> API Keys
+3. Create and copy your API key
+
+### 2. Configure environment variable
+
+Linux/macOS:
+```bash
+export FELO_API_KEY="your-api-key-here"
+```
+
+Windows PowerShell:
+```powershell
+$env:FELO_API_KEY="your-api-key-here"
+```
+
+## How to Execute
+
+Use Bash tool commands and follow this workflow exactly.
+
+### Step 1: Precheck API key
+
+```bash
+if [ -z "$FELO_API_KEY" ]; then
+  echo "ERROR: FELO_API_KEY not set"
+  exit 1
+fi
+```
+
+If key is missing, stop and return setup instructions.
+
+### Step 2: Run Node Script (create + poll)
+
+Use the bundled script (no `jq` dependency):
+
+```bash
+node felo-slides/scripts/run_ppt_task.mjs \
+  --query "USER_PROMPT_HERE" \
+  --interval 10 \
+  --max-wait 1800 \
+  --timeout 60
+```
+
+Script behavior:
+- Creates task via `POST https://openapi.felo.ai/v2/ppts`
+- Polls via `GET https://openapi.felo.ai/v2/tasks/{task_id}/historical`
+- Treats `COMPLETED`/`SUCCESS` as success terminal (case-insensitive)
+- Treats `FAILED`/`ERROR` as failure terminal
+- Stops polling immediately on terminal status
+- Prints `ppt_url` on success (fallback: `live_doc_url`)
+
+Optional debug output:
+
+```bash
+node felo-slides/scripts/run_ppt_task.mjs \
+  --query "USER_PROMPT_HERE" \
+  --interval 10 \
+  --max-wait 1800 \
+  --json \
+  --verbose
+```
+
+This outputs structured JSON including:
+- `task_id`
+- `task_status`
+- `ppt_url`
+- `live_doc_url`
+- `livedoc_short_id`
+- `ppt_business_id`
+
+### Step 4: Return structured result
+
+On success, return:
+- `ppt_url` immediately (script default output, fallback `live_doc_url`)
+- if `--json` is used, also include `task_id`, terminal status, and optional metadata
+
+## Output Format
+
+Use this response structure:
+
+```markdown
+## PPT Generation Result
+- Task ID: <task_id>
+- Status: <status>
+- PPT URL: <ppt_url>
+- Live Doc URL: <live_doc_url or N/A>
+
+## Notes
+- livedoc_short_id: <value or N/A>
+- ppt_business_id: <value or N/A>
+```
+
+Error format:
+
+```markdown
+## PPT Generation Failed
+- Error Type: <error code or category>
+- Message: <readable message>
+- Suggested Action: <next step>
+```
+
+## Error Handling
+
+Known API error codes:
+- `INVALID_API_KEY` (401): key invalid or revoked
+- `PPT_TASK_CREATE_FAILED` (502): create task downstream failed
+- `PPT_TASK_QUERY_FAILED` (502): query task downstream failed
+
+Timeout handling:
+- If timeout reached, return last known status and instruct user to retry later
+- Include `task_id` so user can query again
+
+## Important Notes
+
+- Always execute this skill when user intent is slide generation.
+- Always return `task_id` so follow-up queries can continue from the same task.
+- Do not claim completion without a terminal status.
+- Keep API calls minimal: create once, then poll.
+
+## References
+
+- [Felo PPT Task API](https://openapi.felo.ai/docs/api-reference/v2/ppt-tasks.html)
+- [Felo Open Platform](https://openapi.felo.ai/docs/)
+