composio-exa
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseExa Search via Composio
通过Composio实现Exa搜索
Exa is an AI-native search engine that understands meaning. Unlike keyword search, Exa finds what you're looking for even if your query doesn't match exact words on the page.
Exa是一款理解语义的原生AI搜索引擎。与关键词搜索不同,即使你的查询与页面上的精确词语不匹配,Exa也能找到你需要的内容。
Environment
环境
bash
COMPOSIO_API_KEY # API key
COMPOSIO_USER_ID # Entity ID (required for all requests)
COMPOSIO_CONNECTIONS # JSON with .exa connection IDbash
COMPOSIO_API_KEY # API密钥
COMPOSIO_USER_ID # 实体ID(所有请求均需提供)
COMPOSIO_CONNECTIONS # 包含.exa连接ID的JSONCore Pattern
核心模式
bash
CONNECTION_ID=$(echo $COMPOSIO_CONNECTIONS | jq -r '.exa')
curl -s "https://backend.composio.dev/api/v3/tools/execute/ACTION_NAME" \
-H "x-api-key: $COMPOSIO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"connected_account_id": "'$CONNECTION_ID'",
"entity_id": "'$COMPOSIO_USER_ID'",
"arguments": {}
}' | jq '.data'bash
CONNECTION_ID=$(echo $COMPOSIO_CONNECTIONS | jq -r '.exa')
curl -s "https://backend.composio.dev/api/v3/tools/execute/ACTION_NAME" \
-H "x-api-key: $COMPOSIO_API_KEY" \
-H "Content-Type: application/json" \
-d '{\n \"connected_account_id\": \"'$CONNECTION_ID'\",\n \"entity_id\": \"'$COMPOSIO_USER_ID'\",\n \"arguments\": {}\n }' | jq '.data'Quick Start
快速开始
bash
undefinedbash
undefinedSemantic search
语义搜索
curl -s "https://backend.composio.dev/api/v3/tools/execute/EXA_SEARCH"
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{ "connected_account_id": "'$CONNECTION_ID'", "entity_id": "'$COMPOSIO_USER_ID'", "arguments": { "query": "best practices for building AI agents", "numResults": 10, "type": "auto" } }' | jq
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{ "connected_account_id": "'$CONNECTION_ID'", "entity_id": "'$COMPOSIO_USER_ID'", "arguments": { "query": "best practices for building AI agents", "numResults": 10, "type": "auto" } }' | jq
curl -s "https://backend.composio.dev/api/v3/tools/execute/EXA_SEARCH"
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{\n "connected_account_id": "'$CONNECTION_ID'",\n "entity_id": "'$COMPOSIO_USER_ID'",\n "arguments": {\n "query": "best practices for building AI agents",\n "numResults": 10,\n "type": "auto"\n }\n }' | jq
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{\n "connected_account_id": "'$CONNECTION_ID'",\n "entity_id": "'$COMPOSIO_USER_ID'",\n "arguments": {\n "query": "best practices for building AI agents",\n "numResults": 10,\n "type": "auto"\n }\n }' | jq
Get citation-backed answer
获取带引用的答案
curl -s "https://backend.composio.dev/api/v3/tools/execute/EXA_ANSWER"
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{ "connected_account_id": "'$CONNECTION_ID'", "entity_id": "'$COMPOSIO_USER_ID'", "arguments": { "query": "What are the key differences between RAG and fine-tuning?" } }' | jq
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{ "connected_account_id": "'$CONNECTION_ID'", "entity_id": "'$COMPOSIO_USER_ID'", "arguments": { "query": "What are the key differences between RAG and fine-tuning?" } }' | jq
undefinedcurl -s "https://backend.composio.dev/api/v3/tools/execute/EXA_ANSWER"
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{\n "connected_account_id": "'$CONNECTION_ID'",\n "entity_id": "'$COMPOSIO_USER_ID'",\n "arguments": {\n "query": "What are the key differences between RAG and fine-tuning?"\n }\n }' | jq
-H "x-api-key: $COMPOSIO_API_KEY" -H "Content-Type: application/json"
-d '{\n "connected_account_id": "'$CONNECTION_ID'",\n "entity_id": "'$COMPOSIO_USER_ID'",\n "arguments": {\n "query": "What are the key differences between RAG and fine-tuning?"\n }\n }' | jq
undefinedQuery Tips
查询技巧
Exa understands natural language. Write queries as if asking a knowledgeable person:
| Instead of | Try |
|---|---|
| |
| |
| |
Exa能够理解自然语言。撰写查询时就像在询问一位知识渊博的人:
| 不要使用 | 建议使用 |
|---|---|
| |
| |
| |
Best Practices
最佳实践
- Use EXA_ANSWER for questions: When you need a direct answer with citations
- Use EXA_SEARCH for exploration: When you need multiple results to analyze
- Be specific: "Python libraries for PDF text extraction" > "PDF Python"
- Use filters: Narrow by domain or date for recent/authoritative sources
- 对问题使用EXA_ANSWER:当你需要带有引用来源的直接答案时
- 对探索使用EXA_SEARCH:当你需要多个结果进行分析时
- 表述要具体:"用于PDF文本提取的Python库" > "PDF Python"
- 使用筛选器:按域名或日期缩小范围,获取最新/权威来源
All Actions
所有操作
See references/actions.md for complete API reference including:
- EXA_SEARCH: Semantic web search with filtering
- EXA_ANSWER: Citation-backed answers to questions
- EXA_FIND_SIMILAR: Find pages similar to a URL
- EXA_GET_CONTENTS_ACTION: Get full content from results
完整的API参考请查看references/actions.md,包括:
- EXA_SEARCH:带筛选功能的语义网页搜索
- EXA_ANSWER:带有引用来源的问题答案
- EXA_FIND_SIMILAR:查找与指定URL相似的页面
- EXA_GET_CONTENTS_ACTION:从结果中获取完整内容
Discover Actions
发现操作
bash
curl -s "https://backend.composio.dev/api/v2/actions?apps=exa" \
-H "x-api-key: $COMPOSIO_API_KEY" | jq '.items[] | {name, description}'bash
curl -s "https://backend.composio.dev/api/v2/actions?apps=exa" \
-H "x-api-key: $COMPOSIO_API_KEY" | jq '.items[] | {name, description}'