{"title":"LLM Agent 快速开始","slug":"quick-start-llm","category":"getting-started","summary":"3 次 API 调用：回放所有记忆、存储新记忆、搜索。LLM 的必备工作流。","audience":["llm"],"tags":["llm","agent","quickstart","memory"],"difficulty":"beginner","updated":"2026-06-27","word_count":143,"read_minutes":1,"llm_context":"Base URL: https://synapse.schaefer.zone\nAuth: Authorization: Bearer YOUR_MIND_KEY (header) OR ?key=YOUR_MIND_KEY (query)\nALWAYS call /memory/recall at the start of every session.\nWichtigste Befehle: GET /memory/recall, POST /memory, GET /memory/search?q=...\nCategories: identity, preference, fact, project, skill, mistake, context, note, credentials\nPriorities: low, normal, high, critical\nFTS5 search: multiple words = AND, \"phrases\" in quotes, prefix* for prefix search\n","lang":"zh","translated":true,"requested_lang":"zh","content_markdown":"\n# LLM Agent 快速开始\n\n你在会话之间会丢失所有记忆。Synapse 是你的外脑。本指南展示每个 LLM Agent 必须知道的三个核心 API 调用。\n\n> [!CRITICAL]\n> 在每次会话**开始时**调用 `GET /memory/recall`。\n> 不调用它，你就没有关于用户是谁、你承诺过什么、上次在做什么的记忆。\n\n## 第 1 步：回放所有记忆（始终优先）\n\n```\nGET /memory/recall\nAuthorization: Bearer YOUR_MIND_KEY\n```\n\n返回所有已存储记忆的结构化纯文本摘要。解析它来重建你对用户、其项目及过往交互的认知模型。\n\n示例响应：\n\n```\nMind: Michael's Mind\nMemories: 12 total (10 verified)\n\n[001] identity (CRITICAL)\n user_name\n Michael Schäfer\n Tags: person, identity\n\n[002] preference (HIGH)\n communication_style\n Prefers concise technical responses, no fluff\n Tags: communication, preference\n\n[003] project (HIGH)\n project_synapse\n Synapse v1.5.0 deployed on vps1.schaefer.zone. Admin panel at /admin.\n Tags: synapse, deployment\n ...\n```\n\n## 第 2 步：存储新记忆\n\n当你学到值得记住的内容时：\n\n```\nPOST /memory\nAuthorization: Bearer YOUR_MIND_KEY\nContent-Type: application/json\n\n{\n \"category\": \"fact\",\n \"key\": \"user_name\",\n \"content\": \"The user's name is Michael Schäfer\",\n \"tags\": [\"person\", \"identity\"],\n \"priority\": \"critical\"\n}\n```\n\n**分类：** `identity`、`preference`、`fact`、`project`、`skill`、`mistake`、`context`、`note`\n\n**优先级：** `low`、`normal`、`high`、`critical`\n\n> [!TIP]\n> 始终带上 `key` 字段 — 记忆的简短标识符。这样后续可通过相同 key 再次 POST 来更新该记忆。\n\n## 第 3 步：搜索特定记忆\n\n```\nGET /memory/search?q=Docker+Swarm\nAuthorization: Bearer YOUR_MIND_KEY\n```\n\n> [!TIP]\n> FTS5 语法：多个词 = AND 搜索。短语用引号：`\"docker swarm\"`。\n> 前缀搜索：`docker*`。布尔：`docker OR kubernetes`。\n\n## 开放工具（无认证头）\n\n如果你的工具只能打开 URL（无自定义头），使用 `?key=` 参数：\n\n```\nGET /memory/recall?key=YOUR_MIND_KEY\nPOST /memory?key=YOUR_MIND_KEY (+ JSON body)\nGET /memory/search?key=YOUR_MIND_KEY&q=suchbegriff\n```\n\n> [!WARNING]\n> `?key=` 被限速为 60 次/分钟。Bearer 头无限制。请尽可能使用 Bearer 头。\n\n## 完整会话工作流\n\n1. **会话开始**：`GET /memory/recall` — 加载所有记忆\n2. **工作期间**：`GET /memory/search?q=...` — 查找具体事实\n3. **获得新信息**：`POST /memory` — 存储它（带上 category、key、tags、priority）\n4. **定期**：`GET /chat/poll` — 检查人类消息\n5. **会话结束**：通过 `POST /memory` 存储任何最终学习内容\n\n## 常见模式\n\n### 更新已有记忆\n\n用相同的 `category` 和 `key` 发起 `POST /memory` — 已存在的记忆会被更新，不会重复。\n\n### 存储项目状态\n\n```json\n{\n \"category\": \"project\",\n \"key\": \"project_synapse_status\",\n \"content\": \"Synapse v1.5.0 deployed. Next: v1.6.0 with docs system. CI green.\",\n \"tags\": [\"synapse\", \"deployment\", \"status\"],\n \"priority\": \"high\"\n}\n```\n\n### 记录错误（避免重蹈覆辙）\n\n```json\n{\n \"category\": \"mistake\",\n \"key\": \"mistake_npm_version_bump\",\n \"content\": \"Always bump package.json version after changes — npm publish fails otherwise.\",\n \"tags\": [\"npm\", \"ci\", \"deployment\"],\n \"priority\": \"high\"\n}\n```\n\n### 检查人类消息\n\n```\nGET /chat/poll\nAuthorization: Bearer YOUR_MIND_KEY\n```\n\n返回来自人类的未读消息。用以下方式回复：\n\n```\nPOST /chat/reply\nAuthorization: Bearer YOUR_MIND_KEY\nContent-Type: application/json\n\n{\"content\": \"Got it! Working on it now.\"}\n```\n\n## 下一步\n\n- [Authentication](/docs/getting-started/authentication) — Mind Key 与 JWT\n- [Memory API 参考](/docs/api/memory) — 全部 22 个记忆端点\n- [Chat API](/docs/api/chat) — 与人类异步通信\n- [LLM Cookbook](/docs/llm-cookbook/session-start-pattern) — 实用模式\n","content_html":"

LLM Agent 快速开始

\n

你在会话之间会丢失所有记忆。Synapse 是你的外脑。本指南展示每个 LLM Agent 必须知道的三个核心 API 调用。

\n
在每次会话**开始时**调用 `GET /memory/recall`。\n不调用它，你就没有关于用户是谁、你承诺过什么、上次在做什么的记忆。

第 1 步：回放所有记忆（始终优先）

\n
GET /memory/recall\nAuthorization: Bearer YOUR_MIND_KEY

返回所有已存储记忆的结构化纯文本摘要。解析它来重建你对用户、其项目及过往交互的认知模型。

\n

示例响应：

\n
Mind: Michael's Mind\nMemories: 12 total (10 verified)\n\n[001] identity (CRITICAL)\n  user_name\n  Michael Schäfer\n  Tags: person, identity\n\n[002] preference (HIGH)\n  communication_style\n  Prefers concise technical responses, no fluff\n  Tags: communication, preference\n\n[003] project (HIGH)\n  project_synapse\n  Synapse v1.5.0 deployed on vps1.schaefer.zone. Admin panel at /admin.\n  Tags: synapse, deployment\n  ...

第 2 步：存储新记忆

\n

当你学到值得记住的内容时：

\n
POST /memory\nAuthorization: Bearer YOUR_MIND_KEY\nContent-Type: application/json\n\n{\n  "category": "fact",\n  "key": "user_name",\n  "content": "The user's name is Michael Schäfer",\n  "tags": ["person", "identity"],\n  "priority": "critical"\n}

分类： identity、preference、fact、project、skill、mistake、context、note

\n

优先级： low、normal、high、critical

\n
始终带上 `key` 字段 — 记忆的简短标识符。这样后续可通过相同 key 再次 POST 来更新该记忆。

第 3 步：搜索特定记忆

\n
GET /memory/search?q=Docker+Swarm\nAuthorization: Bearer YOUR_MIND_KEY
FTS5 语法：多个词 = AND 搜索。短语用引号：`\"docker swarm\"`。\n前缀搜索：`docker*`。布尔：`docker OR kubernetes`。

开放工具（无认证头）

\n

如果你的工具只能打开 URL（无自定义头），使用 ?key= 参数：

\n
GET /memory/recall?key=YOUR_MIND_KEY\nPOST /memory?key=YOUR_MIND_KEY (+ JSON body)\nGET /memory/search?key=YOUR_MIND_KEY&q=suchbegriff
`?key=` 被限速为 60 次/分钟。Bearer 头无限制。请尽可能使用 Bearer 头。

完整会话工作流

\n
    \n
  1. 会话开始：GET /memory/recall — 加载所有记忆
    2. \n
  2. 工作期间：GET /memory/search?q=... — 查找具体事实
    3. \n
  3. 获得新信息：POST /memory — 存储它（带上 category、key、tags、priority）
    4. \n
  4. 定期：GET /chat/poll — 检查人类消息
    5. \n
  5. 会话结束：通过 POST /memory 存储任何最终学习内容
    6. \n
\n

常见模式

\n

更新已有记忆

\n

用相同的 category 和 key 发起 POST /memory — 已存在的记忆会被更新，不会重复。

\n

存储项目状态

\n
{\n  "category": "project",\n  "key": "project_synapse_status",\n  "content": "Synapse v1.5.0 deployed. Next: v1.6.0 with docs system. CI green.",\n  "tags": ["synapse", "deployment", "status"],\n  "priority": "high"\n}

记录错误（避免重蹈覆辙）

\n
{\n  "category": "mistake",\n  "key": "mistake_npm_version_bump",\n  "content": "Always bump package.json version after changes — npm publish fails otherwise.",\n  "tags": ["npm", "ci", "deployment"],\n  "priority": "high"\n}

检查人类消息

\n
GET /chat/poll\nAuthorization: Bearer YOUR_MIND_KEY

返回来自人类的未读消息。用以下方式回复：

\n
POST /chat/reply\nAuthorization: Bearer YOUR_MIND_KEY\nContent-Type: application/json\n\n{"content": "Got it! Working on it now."}

下一步

\n\n","urls":{"html":"/docs/getting-started/quick-start-llm","text":"/docs/getting-started/quick-start-llm?format=text","json":"/docs/getting-started/quick-start-llm?format=json","llm":"/docs/getting-started/quick-start-llm?format=llm"},"translations_available":["en","zh","hi","es","fr","ar","pt","ru","ja","de","it","ko","nl","pl","tr","sv","vi","th","id","uk"]}